Evolutionary Developments

of Today’s Remote Sensing

Radar Technology—Right
From the Telemobiloscope
A review

SAMEDH SACHIN KARI, A AROCKIA BAZIL RAJ ,

AND BALASUBRAMANIAN. K

©SHUTTERSTOCK.COM/JAMESTEOHART

T oday, remote sensing systems/technologies are one of

the most essential requirements for civil and military
sectors for various applications. This review article discuss-
from the telemobiloscope. This review article addresses the
fundamentals of radar sensing techniques, top-level radar
classifications, and revolutionary developments of antenna
es the evolutionary developments of today’s remote sensing technologies for remote sensing applications. The various
radar/optical/electronic warfare (EW) technologies, right techniques available for radar waveform design, a char-
acteristics analysis of it using ambiguity functions (AFs),
Digital Object Identifier 10.1109/MGRS.2023.3329928
pulse compression/stretch processing, a time-frequency
Date of current version: 20 November 2023 (T-F) analysis, and so on are reviewed. The significant

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 2473-2397/23©2023IEEE 67

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 transformations that have happened in radar system engi- were at lower power. In the 1940s, the Radiation Lab at
neering since vacuum tube microwave devices are reported. the Massachusetts Institute of Technology developed an
Contemporary societal applications of radar systems, track- S-band 250-KW (high-power) radar SCR-584 with an an-
ing and guidance radar systems, advanced EW systems, tenna diameter of 2 m based on the gunfire control system
photonics EW systems, and photonics signal processing are [140] for use in WWII. During the same years, the Rus-
reviewed and reported. State-of-the-art optical technolo- sia developed the high-power radar systems RUS-1 and
gies available for today’s remote sensing applications are RUS-2, mass produced them, and used them effectively in
discussed. In addition to these reviews, a comprehensive WWII. In parallel, the United Kingdom’s Royal Air Force
comparative study is performed in terms of available remote developed the first airborne radar, in 1940, known as Air-
sensing systems/technologies, their typical operating fre- borne Interception Mark IV, for use in WWII. Thus, signifi-
quency ranges, potential applications, types of waveforms, cant research happened in the 1930s, and many countries
and so forth, and the quantitative results are reported. had operational radar systems that played a vital role dur-
ing WWII (1939–1945).
INTRODUCTION In the 1930s, sound navigation and ranging-based re-
Beginning with ancient/prehistoric civilizations, humans mote sensing systems were developed based on the prin-
have observed Earth through a bird’s-eye view, i.e., an ciples of radar, mainly for underwater communications and
aerial view [1]. During the evolution of civilizations, re- target detections. Further, subsequently, research led to the
quirements, such as finding available resources, managing/ development of more specialized radars like monopulse
utilizing them appropriately, observing the hazards, setting radar, Doppler radar, and synthetic aperture radar (SAR).
up a suitable defense system, and so on, have demanded SAR was invented in 1951 by Carl Wiley, a mathemati-
remote sensing at those times [2]. These mandatory remote cian working at Goodyear Aerospace Corporation. Devel-
sensing requirements have driven radar technology to great opments in computer technology transformed the image
heights/successes and are now finding various applications processing techniques of images formed by radars [4]. As
in automobiles, biomedical, aerial traffic control, synthetic space technology advanced, radars were carried to space as
aperture geographical imaging, detecting foliage-hidden satellite payloads for continuous sensing/imaging of Earth.
targets, underground subsurface imaging, deep-space sens- Seasat was the first satellite that carried an SAR to space [4].
ing, satellite-based Earth imaging, weather forecasting, as- In the1960s, Hughes Aircraft Company (USA) developed
tronomical radar imaging, and much more. Early attempts the lidar system based on the working principle of radar
at remote sensing/imaging began with the development but with optical wavelength as the carrier [8]. In recent de-
of optical sensors, i.e., cameras, which captured pictures cades, with the development of aircraft, unmanned aerial
in the optical, i.e., visible/infrared (IR) frequency bands. vehicles (UAVs), small UAVs (sUAVs), the use of airborne-
At later stages, to get wider coverage, these optical sensors based remote sensing/imaging with radar/lidar systems has
were attached to kites and balloons; then, images of larger increased dramatically [9].
dimensions were captured from higher-altitude points [1]. The remainder of this review article is organized as fol-
One of the earliest attempts at radio-frequency (RF) de- lows. The “Fundamentals of Radar Sensing Technique”
tection and ranging was the “telemobiloscope,” made by section briefs the fundamentals of radar technology. The
Christian Hülsmeyer in 1904. It was a simple (and also the “Radar System Classification” section gives the classifica-
first) device with an active RF sensor that used radio waves tion of radar systems. The “Remote Sensing Radar Systems
to detect the presence of ships around it. But the telemobi- Based on Operating Environments” section explores radar
loscope neither found the direction nor the distance, it just systems based on different operating environments. The
detected the presence of ships in a limited range [3], [4]. As “Developments in Antenna Technology for Remote Sens-
the radar technology was in its primitive stages during the ing” section explains the developments of antenna for
early 1900s, it did not find much use in WWI (1914–1918). radar technology. The “Radar Waveform Design and Its
Detection of aircrafts was done by passive acoustic sensors Processing Techniques” section describes the radar wave-
during WWI. Acoustic sensors took advantage of the fact form design and signal processing techniques. The “Devel-
that the engines of the airplanes, helicopters, and tanks opments in a Radar Systems’ Engineering Design” section
used during WWI were significantly noisy. However, acous- details the developments of radar system engineering de-
tic sensing technology had its own limitations, like low signs. The “Contemporary Societal Applications of Radar
range, susceptibility to background noise, and dependency Systems” section presents the contemporary applications
on atmospheric conditions/effects [5]. of today’s radar systems. The “Tracking and Guidance Ra-
After WWI, several countries, including the United dars” section reports the tracking and guidance radars. The
Kingdom, Germany, Russia, the United States, and France, “Modern Staring Radar Systems” section narrates modern
started investigating/developing resources to make opera- staring radar systems. The “EW Systems” section briefs EW
tional and efficient radars [6]. Radar, in the early stages systems. The “Photonic EW Systems and Signal Process-
of such development, did not have high-power transmit- ing” section gives the photonic EW systems and photonic
ters. So, the radars used at the beginning (1939) of WWII signal processing techniques. The “Optical Technology for

68 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 Remote Sensing” section describes optical remote sensing where a, h, i, and k are the radius, height, and tilt angle of a
technology. The “Characteristics Comparison of Today’s cylinder target and wave number, respectively.
Remote Sensing Systems/Technologies” section gives the
v = 4r a
ab k2
characteristics comparison of today’s available remote sens- (5)
m
ing systems/technologies in a table form. Finally, the “Con-
clusion” section gives the conclusion. where a and b is the length and height of a flat plate, respec-
tively, and ab has to be greater than the m.
FUNDAMENTALS OF RADAR SENSING TECHNIQUES
v = 8r a
ab k2
This section gives an understanding of different essen- (6)
m
tial, fundamental concepts associated with radar sens-
ing techniques. where a and b is the height and length of a dihedral corner
reflector, respectively, and ab has to be greater than the m.
RADAR CROSS SECTION
v = 12r a
ab k2
Radar is usually an active RF sensor. It produces electro- (7)
m
magnetic (EM) signals that travel to the target and get re-
flected by it. The signal incident on the target is actually where a and b is the height and length of a trihedral rect-
scattered in all directions, and a part of it [based on the angular corner reflector, and ab has to be greater than the
radar cross section (RCS) of the target] travels back to the m. The RCS of complex-shaped targets like ground terrain,
receiving antenna. By detecting, processing, and analyzing aircraft, ship, mountains, hills, tanks, forest, and so forth
the received wave, it is possible to remotely sense several is equal to the combined scattering phenomena of all the
parameters, such as position, velocity, structural dimen- simple-shape scatterers for which the solution can be ob-
sions, behaviors, and so on associated with the target. tained by using the well-known method of moments, or
Thus, the RCS of the target, which is actually the ratio of finite-difference time-domain solvers [12]. Further, the
intensities of received and incident signals, is one of the value of the RCS has to be fixed while designing/devel-
most important parameters in radar systems. General- oping radar specifications/systems in the case of known
form RCS, v, is [10], [11] targets (e.g., a monostatic collision-avoidance radar sys-
tem), whereas in the case of unknown targets (e.g., enemy
E 2r sUAVs), the RCS of them have to be above the threshold/
v= 4rR 2 (1)
E 2i expected value for which the radar is designed for their
detection. In addition to this, today, targets are designed
where E r is the reflected electric field, E i is the incident with stealth capabilities, i.e., altering geometrical struc-
electric field, and R is the distance between the transmit- tures, applying RF-absorbing paints, and designing with
ting radar and the target. Thus, RCS characteristics, i.e., composite RF-absorbing materials, to reduce the RCS so
scattering properties of the structures, of the target com- as not to get detected by any radar system. However, such
pletely depend on the RF scattered-field equation as stealth targets can be detected while widely chirping the
radar frequency because the RCS depends on it, as in (2)
jk 0 E 0 and (3)–(7). Thus, the scattering property/characteristics
E s (r) = - 4rr e -jk0 r ##S lit
2nt (r l ) # (kt i # ut ) e j (k - k ) r l d 2 r l (2)
s i

of the target’s structures can increase or decrease the RCS,

which in turn makes the targets either visible or not vis-
where k 0 is the wave number; E 0 and ut are the magnitude ible by the radar system.
and polarization direction of incident electric field, respec-
tively; nt (r l ) is the outward unit vector of the illuminated RF RADIATORS
object’s surface; r l is the origin to any point on the illumi- Monostatic radar has a single antenna, which is used for
nated surface (S lit) of the object; and kt i, k s, and k i repre- both transmission and reception of RF radiation. It uses a
sent the incident wave vectors, and the derived analytical duplexer that acts as a router between the transmit and re-
models from this equation for computation of the RCS of ceive chains, as shown in Figure 1(a). As only one antenna
various simple-shaped targets sphere, cylinder, flat plate,
dihedral corner reflector, and trihedral rectangular corner
reflector are Transmitter Transmitter

v = ra 2 (3) Duplexer

where a is the radius of the sphere, and a has to be greater Receiver Receiver
than the wavelength ^mh of the transmitted radar signal
(a) (b)
2
am cos i sin (kh sin i) FIGURE 1. An antenna-based classification of radars. (a) Monos-
v = 2r (4)
sin 2 i tatic radar and (b) bistatic radar.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 69

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 is used, the gain (G) of transmission ^G T h and reception radar in time T, and c is the speed of the EM wave, then
^ G R h chains are equal, i.e., G T = G R = G. With the radar the range of a scattering point target is given by
equation, we can get the range of the target in the mono-
cT
static radar as [14] R = 2 . (12)
1

R max = < F (8)

PT G 2 m 2 v 4 However, moving target indicator (MTI) radar can
(4r) 3 S min detect moving targets by taking advantage of the phase
shift, i.e., the Doppler frequency, that causes a frequen-
where S min is receiver sensitivity, i.e., the minimum receiv- cy shift on the reflected EM wave in accordance with
able power, and PT is the transmission power [13]. the velocity of the target [20]. MTI radar uses low pulse
Bistatic radar, on the other hand, has two separate an- repetition frequency (PRF) to avoid range ambiguities.
tennas dedicated to simultaneous transmission and recep- This limitation of unmodulated CW radar is solved by
tion of RF waves, as shown in Figure 1(b). Usually, bistatic developing frequency-modulated CW (FMCW) radar
radar has the transmission and reception antennas at a sig- [21], which calculates both the speed and distance of
nificant distance. Hence, in general, the transmitter and the the target [12]. The range computation in FMCW radar
receiver antennas will have separate ranges. The radar equa- is given by [15]
tion for a bistatic radar system is
c | Df |
R= (13)
2 df
PT G T G R m v 2 dt
PR = (9)
(4r) 3 R 2T R 2R
where f is the radar frequency, and t is the delay time. Pres-
where R T and R R are the transmission and reception rang- ently, due to its lower peak power, FMCW radar has become
es, respectively [14]. more suitable for short-range applications. However, some
research has been done to reduce the mutual EM signal
PT G T G R m 2 v coupling among the antennas in the FMCW radar system
PR = . (10)
(4r) 3 R 4 so that the high-power radiation by FMCW radar is able to
operate it for long-range applications/missions.
The total power of the receiving antenna in a bistatic ra-
dar when it is near the transmitting antenna, and assuming DETECTION DISPLAYS
that the target is a point scatterer, (9) becomes (10), where Right from the initial periods, radars have used CRT-
R T = R R = R. based displays. A-scope, B-scope, C-scope, E-scope, and
plan-position indicators (PPI) are a few of the radar dis-
EM SIGNALS plays types. A-scope is a basic radar display that plots the
Continuous-wave (CW) radars generally use the bistatic ra- amplitude-versus-range readings of the radar system. It
dar in which one antenna is dedicated to transmit the EM provides just the range data and not the direction data
signals and other one is dedicated to receive the EM sig- of the target. The process of data acquisition for an A-
nals. An unmodulated CW radar is known as Doppler radar scope display is known as A-scan. B-scope is another type
because it is capable of detecting only the velocity v of the of display and shows the plot for range versus azimuth.
target and not the distance as B-scan is the process of data acquisition for B-scope.
E-scope is similar to B-scope but it displays the range-
fd m versus-against elevation instead of azimuth. C-scope is
v = 2 cos (i)(11)
another radar output display that is circular and plots the
perspective view, giving the azimuth-versus-elevation
where fd is the Doppler shift between the transmitted plot of the target. The process of data acquisition for a
and received signals, and i is the angle the velocity vec- C-scope display is known as C-scan. The PPI display is a
tor makes with the radar. Pulse radar generates pulses of very popular display that maps the range and angle on a
EM waves at a repetition rate, known as pulse-repetition polar-coordinate map. It plots all 360° around the radar
time (PRT), keeping the pulse on for a fixed time duration, up to a limited range. A sector PPI is a special case of PPI
known as pulsewidth ^x h . They usually use the monostatic that does not map the entire 360° circular plot, just a
antenna, thus, EM is radiated for the period of x while particular sector of it.
the EM waves are received by the same antenna for the pe-
riod of PRT- x. Thus, the time of transmission is typically RADAR SYSTEM CLASSIFICATION
much less when compared to radar-listening time. Basic Classifications of radar systems are based on the following
pulse radar is capable of detecting only the range of the general criteria and briefly reported in this section: 1) the
target by calculating the time taken by the pulse to travel technology used to develop it, 2) frequency ranges used to
to the target and back. If the pulse returns back to the operate it, and 3) applications for which it is developed.

70 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 TECHNOLOGY-BASED CLASSIFICATION APPLICATION-BASED CLASSIFICATION
Radar technologies are mainly classified into two catego- Radars are used for various applications and missions,
ries on the basis of 1) the antennas used and 2) working which include advanced driver assistance systems (ADASs)
principles/characteristics. Based on the antennas, radars in cars, biomedical radars for health purposes, weather ra-
are of two types: monostatic and bistatic [14]. Based on dars to forecast the weather, SAR and inverse SAR (ISAR)
the working characteristics, radars are classified into two for imaging of Earth and penetration depth, missile guid-
types: CW and pulse, as shown in Figure 2. The RF con- ance radars, ground-penetrating radar (GPR), foliage pen-
figurations based on pulse radars gives the MTI and MT etration (FOPEN) radar, air traffic management radars,
displays, which support the detection of slow-/fast-mov- and many more. Each application demands some modi-
ing targets [15]. Based on the frequency modulation, CW fications in the RF assembly, and signal processing algo-
radar is classified into two types: linear FMCW (LFMCW) rithms of radars used to suit its purpose. Therefore, radar
and stepped FMCW (SFMCW). The frequency modula- systems are also classified based on the applications, as
tion in CW radars is typically increased chirp, however, shown in Figure 3.
based on applications, sometimes
the chirp can be of a decreased or
triangular profile.
Types of
Radars
FREQUENCY-BASED
CLASSIFICATION
The penetration and reflection phe- Based on Based on
Antenna Characteristic
nomena of EM waves depend on
the dielectric property of the target
and the radar wavelength chosen. Monostatic Bistatic CW Radar Pulse Radar
Radars are used for various applica-
tions, from biomedical technology
to aerial traffic management, where Moving Target Moving Target
LFMCW SFMCW
every application demands a differ- Indicator Display
ent frequency. Hence, it is important
to choose a suitable frequency for an FIGURE 2. A classification of radar based on technology. LFMCW: linear FMCW; SFMCW:
effective application of the radar [2], stepped FMCW.
[16]. IEEE’s and the U.S. Department
of Defense’s index of specifications
and standards have standardized/ TABLE 1. RADAR CLASSIFICATION BASED ON FREQUENCY AND
classified radars into 13 types (e.g., RESPECTIVE APPLICATIONS [18], [19].
L band, X band, W band, and so
SERIAL BAND FREQUENCY
forth) based on the appropriate fre- NUMBER DESIGNATION RANGE TYPICAL APPLICATION
quency ranges, i.e., RF bands, which 1 High 3–30 MHz Over-the-horizon surveillance
are suitable for various applications frequency (HF)
having different targets of various 2 Very HF 30–300 MHz Line-of-sight surveillance, counter stealth, foliage
penetration (FOPEN), ground penetration
dielectric properties/constants [18].
3 Ultra HF 300–1,000 MHz Long-range surveillance, FOPEN
The name of these RF band clas-
4 L 1–2 GHz Long-range surveillance, air traffic control
sifications are listed in Table 1 [19]
5 S 2–4 GHz Terminal air traffic control, long-range weather
with their typical intended applica- observation, airborne early warning
tions. The colloquial (undocument- 6 C 4–8 GHz Weather observation, weapon location
ed) expansions of the letter codes L 7 X 8–12 GHz Short-range tracking, missile guidance, marine
through W are also given in Table 1. radar mapping
As most of the technology was de- 8 Ku 12–18 GHz High-resolution mapping, satellite
altimetry, UAV radar
veloped during WWII, the naming
9 K 18–27 GHz Police radar
scheme (letter codes such as S, C,
10 Ka 27–40 GHz Very short-range tracking, airport
L, X, and so forth) that followed for surveillance
the RF bands was too cryptic to des- 11 V 40–75 GHz Scientific remote sensing
ignate the district frequency bands 12 W 75–110 GHz Automobile cruise control, missile seeker
at which the radar was operated to 13 mm 110–300 GHz Experimental
keep the enemy from knowing it,
L: long; S: short; C: compromise; X: secret during W WII; Ku: under K; K: Kurz; Ka: above Ka; V: Victor;
and these classifications are still fol- W: whiskey; mm: millimeter.
lowed even today.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 71

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 REMOTE SENSING RADAR SYSTEMS BASED ON early warning and control radars, airborne 3D radars, and
OPERATING ENVIRONMENTS so forth form the RF beam exactly along the flying direc-
This section details remote sensing radars that are in- tion or 90° to it in the azimuth plane to accomplish the
stalled at/operated from different environmental loca- airborne method of operations [26]. An airborne radar
tions/conditions such as airborne platforms, spaceborne operation scenario is shown in Figure 4(a), and in several
platforms, ground points, maritime environments, and applications the airborne radar has connectivity with the
over-the-horizon (OTH) regions. satellite or ground station for the transfer of detection infor-
mation of the aerial targets, and/or getting the command/
AIRBORNE RADARS control data/instructions from the control office located at
Radars mounted on aircraft or UAVs are known as airborne a remote point on the ground.
radars. They usually fly at altitudes ranging from 30 m to
more than 2 km above the ground [23]. The RF beam of SPACEBORNE RADARS
airborne radars does not usually face the ground vertically Spaceborne radars are carried by satellites to space/deep
downward, but rather with a tilted angle, which is known space as payloads to gather data corresponding to the sur-
as a side-looking radar. However, certain applications, like face of the earth or its surrounding atmosphere. Space-
snowfall measurement, radio altimeter, and so on, use a ra- borne radars are functionally similar to airborne radars
dar’s RF beam pointed vertically downward [24]. The radars except for their operation altitudes. The satellites that
that are not side looking have range dependence of clutter. carry spaceborne radars usually orbit at an altitude of
The effects of clutter in airborne radars can be effectively 400 km [27]. By being at such high altitudes, as shown in
handled/canceled using three well-known techniques: a Figure 4(b), they normally offer low resolutions but wider
frequency-diverse array; the multiple-input, multiple-out- footprint/coverage. However, in general, the resolution is
put (MIMO) method, and space–time adaptive processing improved in spaceborne radars, providing the advantag-
(STAP) [25]. Also, operating airborne radars at a medium es of polarization, i.e., the final radar-scanned image is
PRF is preferred when the clutter density/spectrum is more, constructed by combining the data individually collected
whereas in the case of less, low-/high-PRF is preferred. SAR, with the horizontal and vertical polarizations. Particular-
ground surveillance radar, oil spill detection radar, FOPEN ly, in spaceborne radars, altering the antenna pattern and
radar, and so on are the most popular airborne radars. The changing the signal processing techniques are the main
airborne radars of specific missions such as defense radars, methods used for clutter rejection [28]. Spaceborne radars
are mostly used for observing Earth’s atmosphere, weath-
er changes, oil spills on the sea surface, topographical
changes, ocean changes, and so forth [29]. In some cases,
Applications
of Radars spaceborne radars have a direct (otherwise via intermedi-
ate satellites) microwave communication link to ground
Automobile Biomedical Weather stations for the transfer of detection data and/or getting
Radar Radar Radar
the control information.
Air Traffic
SAR ISAR
Management Radar GROUND-BASED RADARS
The radars mounted on the ground surface for terrestrial-
FOPEN Oil Spill Detection
GPR related/based operations, as shown in Figure 4(c), are
Radar Radar
known as ground based. Several radars, including biomedi-
FIGURE 3. A classification of radar based on the application. cal radars, automotive radars, air traffic control radars used

(d)

(a) (b) (c) (e)

FIGURE 4. A classification of radars based on their operating environment. (a) Airborne radar, (b) spaceborne radar, (c) ground-based
radar, (d) maritime radar, and (e) OTH radar.

72 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 at airports, ground-based defense radars, subsurface imag- ionosphere [see Figure 4(e)] toward the targets, and the
ing radar, border security radar, ground weather forecasting scattered EM wave comes back in the same path to the re-
radar, rainfall detection/prediction radars, and many more, ceiving antenna of the OTH radar. The spatial-separation
are operated from the ground surface. However, another distance of both the transmitting and receiving antennas
class of ground-based radars is interferometry based, which is in the order several of several kilometers to avoid the
works based on the principle of interferometry, i.e., phase mutual EM coupling/induc-
difference of waves, which are popular mainly for structural tions among them [32]. OTH
health monitoring. Structural health monitoring includes radars are generally operated
SPACEBORNE RADARS ARE
structural (landslides, landforms, building, dams, bridges, in the HF or very HF (VHF)
and so forth) health monitoring. Ground-based radars are ranges. Surface-wave OTH CARRIED BY SATELLITES TO
mostly used for short-range applications due to Earth’s cur- radars cover large ranges by SPACE/DEEP SPACE AS
vatures, and nonavailability of line of sight (LoS) in most traveling the EM waves along PAYLOADS TO GATHER THE
of their applications. Based on the mission requirements, the surface of the earth or DATA CORRESPONDING TO
ground-based radars are mounted on heavy-load-carrying sea by using the lower fre- THE SURFACE OF THE EARTH
mobile radar trucks, which are moved for the purpose of quencies in the same range OR ITS SURROUNDING
surveillance of specified sensitive regions. The main issues [31]. Clutter for OTH radars
ATMOSPHERE.
with ground-based radars are ground reflections, slow-/ is handled by using spatial
fast-moving ground clutters, building reflections, wind- processing techniques like
velocity profile circulations, and dense fog circulations, slow-time MIMO STAP [33].
which are effectively handled by applying dc-blocking fil- Because the operation range of OTH radar is very high and
ters, clutter mitigations techniques, and signal-to-clutter also via an NLoS path, the detection and localization of
ratio (SCR)-improving algorithms. In the case of mobile the OTH target is very difficult.
radar platforms, the own (i.e., platform/vehicle) Doppler
nullification technique has to be applied in ground-based DEVELOPMENTS IN ANTENNA
radar signal processing to accurately determine the velocity TECHNOLOGY FOR REMOTE SENSING
of the targets canceling the own Doppler introduced by the The evolution and development of antenna technology for
own platform motions. remote sensing is reviewed in this section.

MARITIME RADARS MONOPOLE ANTENNAS

Maritime radars are used for maritime applications in sea A monopole antenna is a linear RF antenna with a length
and ocean regions, which are also known as seaborne radars of m/2, which is in the form of a straight rod with a ground
and, normally installed either on seashores or in ships, as plane used to transmit as well as receive the RF signals [36]
shown in Figure 4(d). Maritime radars are severely affected [see Figure 5(a)] in the initial remote sensing radars. How-
by the sea clutter sourced from sea waves. The effects of ever, monopole antennas have also been realized in planar
sea clutter are more impactful when detecting smaller ves- forms with circular or elliptical shapes [37]. These planar
sels, which invites dual-high-frequency (HF) surface-wave monopole antennas follow an omnidirectional radiation
transmission to significantly overcome this problem [31]. pattern, like the straight rod. They are wide-bandwidth an-
Seashore-based operations of maritime radars are mostly tennas that cover a large range of frequencies, but with less
used to detect and spatially localize marine targets (boats, gain, and have not become popular in the latter days of re-
ships, vessels, sea landmass, buoys, and so on) based on mote sensing applications due to their large size of ground
their applications within their coverage/footprint. The mar- plane/plate [36].
itime radars installed in ships are mainly for maritime sur-
veillance purposes. In both cases, due to the heaving nature DIPOLE ANTENNAS
of the targets, maritime radars can easily form ISAR images A dipole antenna is the simplest and most widely used form
of the targets, e.g., for ships, which are then passed into the of antenna for remote sensing in the initial periods of re-
suitable classification algorithms for the purpose of accu- mote sensing applications. The first dipole antenna used to
rate detection and classification of the targets [30]. transmit and receive radio waves was developed by Heinrich
Hertz during 1886 and 1887 [34]. This dipole antenna con-
OTH RADARS sists of a straight wire of m/2 dipole with a spark gap at the
OTH radars, as the name suggests, have a very long range, center, as shown in Figure 5(b) [34], [35]. When a sinusoi-
up to a few thousand of km, coverage which is beyond dal input is applied/received, the antenna transmits/receives
their horizon. OTH radar is capable of detecting the far- EM waves of frequency (f) equal to f = c/2L, where L is the
thest targets of non-LoS (NLoS), which are otherwise length of the dipole antenna. Several variants of this dipole
nearly impossible. The two types of OTH radars are antenna, like the biconical dipole antenna, folded dipole an-
skywave and surface wave. Skywave radar transmits EM tenna, L-loop antenna, and so forth have been designed and
waves into the atmosphere, which gets reflected by the used for remote sensing purposes in subsequent years [35].

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 73

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 CASSEGRAIN ANTENNAS and S is the separation between two turns, even at low-
The cassegrain antenna is one that has been designed with frequency bands.
two parabolic surfaces, as shown in Figure 5(c). The feed is
on the larger primary reflector, which is behind the smaller MICROSTRIP PATCH ANTENNAS
secondary reflector. The signals received by the antenna are The microstrip patch antenna is a planar-printed one, and
reflected twice before they reach the feed, as shown in Fig- the first monostrip antenna was developed during the
ure 5(c) [38]. The same amount of reflections happen when the 1950s, but it was put to significant use only after the 1980s
feed transmits the signal. This provides several advantages, [42]. It contains a dielectric substrate over a grounded plane.
like less backscattering than a simple parabolic antenna; The dielectric plate has metallic patches printed on it [see
the possibility for mechanical RF beam steering; increas- Figure 5(e)], which is used as the antenna for transmission
ing the transmission gain (G) as G = 6 (D/m) 2, where D is and reception [36] of specified RF bands. The resonant fre-
the diameter of the primary reflector; and the possibility of quency of a microstrip patch-array antenna is given by [42]
getting a focal length greater than the physical length [39].

` mhr j + ` nLr j + a W k (14)

Cassegrain antennas are also designed in multiple variants, 1 2 2 pr 2
(fr) mnp =
have become popular, and are being effectively utilized for 2r nf
various remote sensing, television, satellite broadcasting,
and mobile communication applications, even today. where m, n, and p define the mode of resonance; h is the
height of the substrate; L is the length of the patch; W is the
HELICAL/SPIRAL ANTENNAS width of the patch; and n and f represent the permeability
As the name suggests, helical and spiral antennas are and permittivity of the substrate, respectively. The patch-ar-
two types of antennas that are helical or spiral shaped, as ray antenna became popular for almost all the RF bands due
shown in Figure 5(d) [40], [41]. A helical antenna is made to its light weight, simple design, and fabrication process.
up of wire in a helical shape, with a fixed diameter and dis- Patch-array antennas can be 1D or 2D, with single- or mul-
tance between the turns. An electric field is applied at its RF tiple-RF feeds based on the application. Today, 2D patch-
feed and the radiation comes from the entire helical/spiral array antennas are very popular for remote sensing with
structures. One mode of radiation is perpendicular to the multiple-RF feeds that support electronically steered radar
helical axis (known as normal mode), while the other is beam directions with appropriate phase-shifter networks.
along the helical axis (known as beam radiation), both of
which are possible in helical/spiral antennas [40]. A spiral RFID TAGS
antenna can be of various configurations, like logarithmic, RFID is a basic idea for marking objects with unique RF tags.
Archimedean, or rectangular. Photo etching of the desired These tags contain transponders, which, upon reception of a
spiral on a copper-clad dielectric base material gives the signal, transmit a coded signal, and RFID tags are normally in
different types of spiral antennas. Dual-arm spirals have flexible sheets, as shown in Figure 5(f). The transmitted sig-
better properties and better radiation symmetry. These an- nal contains a unique identification number/code allotted to
tennas are capable of transmitting broadband, circularly the respective object. There are two types of RFID tags avail-
polarized signal [41]. These antennas support RF radiations able: active and passive [43]. Active tags have their own pow-
with increased gain G = 6.2C 2 NS/m 3 = 6.2C 2 NSf 3 /c 3, er source, which is usually the onboard battery. On the other
where C is the circumference of the antenna, c is the veloc- hand, passive tags are powered by a signal from an external
ity of the EM waves in the air, N is the number of turns, reader, known as an RF exciter [43]. These RFID tags are used

(a) (b) (c) (d) (e) (f)

(h)
(g) (i) (j) (k) (l)

FIGURE 5. Developments in antenna technology. (a) A monopole antenna, (b) dipole antenna, (c) cassegrain antenna, (d) helical/spiral
antenna, (e) microstrip patch antenna, (f) an RFID antenna, (g) a slotted-array antenna, (h) phased-array antenna, (i) cavity-backed spiral
antenna, (j) lens antenna, (k) photonic antenna, and (l) plasma antenna.

74 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 for remotely sensing tags that have a lot of information to CAVITY-BACKED SPIRAL ANTENNAS
identify and/or recognize objects. RFIDs are low cost and can A cavity-backed spiral (CBS) antenna is one of the most re-
be quickly installed for establishing a wireless network ar- cent technologies of antennas and represents a modification
chitecture in places where the installation of alternative clas- of spiral and waveguide antennas. The spiral radiator, backing
sical wireless network setups is not possible or risky. Today, cavity, and balun transformer are main components of CBS
various wireless measurements and remote sensing applica- antennas. The spiral antennas are mounted on a closed me-
tions are being done with RFID networks. The gain (G) of tallic waveguide cavity, which makes the bidirectional spiral
these RFID tags is computed as G = 10 log(Pr) - 10 log(Pt) + antenna radiate in only one direction, as shown in Figure 5(i).
10 log (G r) - 20 log (R) - 20 log ( f ) + 10 log (1 - | C t |2) + The spiral radiator is connected to a coaxial cable via the balun
10 log (1 - | C r |2) + 147.6, where Pr is the power transmit- transformer, which is present below the cavity structure [47].
ted by the reader, Pt is the power received by the tag, G r is the The wire thickness of the spirals printed on the surface of the
gain of the reader, R is the range, C t is the reflection coefficient CBS antenna increases as the spiral approaches the outermost
of the tag, and C r is the reflection coefficient of the reader. rings, which in turn is responsible for the high- to low-fre-
quency bands. Due to their compact size, high gain, and mul-
SLOTTED-ARRAY ANTENNAS tiband frequency coverage, CBS antennas are preferred for EW
Well before the introduction of microstrip patch-array anten- applications and long-range remote sensing applications.
nas, waveguide antennas, like horn antennas, were popular
for remote sensing and any high-power radiation applications. LENS ANTENNAS
Slotted-array antennas are of a similar type, except that they Lens antennas, as the name suggests, have a microwave
make slot cuts in the rectangular waveguides for RF radiation, lens, similar to an optical antenna, that focuses the inci-
as shown in Figure 5(g). The most popular slotted structures dent waves on a single point. These lens antennas are wide-
are made either in longitudinal-radiating slots or center-in- ly used for millimeter-wave frequency ranges. A microstrip
clined coupling slots. These antennas were first developed antenna is used as a feed for the lens antenna, as shown
during the 1950s using approximate (as no sophisticated in Figure 5(j). A dielectric lens could be made of poly-
design tools or precise fabrication machines were available) thene with a dielectric constant of approximately 2.55. A
design techniques [44]. Because of their small volume, high rise in gain can be achieved with the antenna after passing
gain, and low side lobes, these antennas give narrow and the signal through the lens [48]. Research on the lens an-
shaped high-gain (G) beam: G = 10 log ((N/2) # (m g /m 0)), tenna design and its optimizations are rapidly coming up
where N is the number of slots, m g is the guide wavelength, and will definitely take a significant place in radar-based
and m 0 is the operational wavelength. Shaping of the beam remote sensing applications. The refractive index of the
is done by several patterned synthesis techniques at later Luneburg lens used for making a lens antenna is given by
stages [45]. Even today, these antennas are used in airport n = 2 - (r/R) 2 , where r is the distance between the point
and maritime remote sensing applications. on the lens, and R is the radius of the lens.

PHASED-ARRAY ANTENNAS PHOTONIC ANTENNAS

In the middle age of remote sensing revolutions, me- Photonic devices use fiber-optic cable (insulators) instead of
chanical beam steering, with electronically controlled conventional conducting lines, i.e., coaxial cables (conduc-
high-torque servomotors, was introduced to increase the tors), which enhances the performance of photonic systems.
number of sensing regions. A mechanical RF beam steering On-the-fly changes in the characteristics of an antenna-like
antenna system is very bulky, heavyweight, and requires gain, the tuning frequency, bandwidth, effective aperture,
a high-voltage power supply; hence, the overall system and so on can be done through photonic devices. These an-
is too complex. The invention of phased-array antennas tennas are called photonically controlled reconfigurable antennas,
significantly transformed beam steering techniques by in- or simply, photonic antennas [49], as shown in Figure 5(k).
creasing the resolution and drastically reducing the overall These antennas work based on photonic switches, which have
weight and design complications. A phased-array anten- several advantages over simple p-i-n diodes or GaAs-based
na is also an array of antennas in which the individual transistors. They are made from photoconductive materials,
antenna(s) transmit the RF signal with a particular phase usually silicon and InGaAs. These materials exhibit a low illu-
difference ^z h, as shown in Figure 5(h). This phase differ- minated-state dc resistance and high dark-state dc resistance.
ence is given by [46] Silicon-based photonic switches can be operated from 2,000 X
when “off” and 6 X when “on,” and the normal operating
2rd sin i
Dz = (15) frequencies of today’s photonic antenna begin at 1 GHz [49].
m

Where d is the separation distance between antennas, and PLASMA ANTENNAS

i is the steered angle of the transmitted beam. This allows Plasma antennas are similar to traditional antennas, where
for the possibility of electronically steering the beam trans- plasma is used for transmission and reception of an RF sig-
mitted by the antenna array [46]. nal instead of the metallic elements in classical antennas,

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 75

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 as shown in Figure 5(l). The plasma is confined within a radar systems, the signal to be transmitted was modulated
dielectric tube in a different orientation so as to get/design and processed after the reception of the radar echo signal
different gaseous plasma antennas in which the filled gas came from the targets. The three main types of radar signal
behaves as a good conductor when the tube is energized, modulations are amplitude modulation, frequency modu-
and discharges when driven at a frequency smaller than lation, and phase modulation, or a combination of these
plasma. Under these conditions, it can transmit or receive modulations through which any specific radar waveform can
EM waves [50]. Plasma antennas have several advantages be designed. Amplitude shift keying is an amplitude on/off
over traditional copper-wire antennas. They can be recon- switching technique used mainly in pulsed-radar systems. An
figured for a different frequency, gain, and bandwidth. This RF signal A cos (2rfc t + i) needs to be transmitted and can
reconfiguration can be done in milliseconds as the switch- be modulated s(t) with PRT pulse data [see Figure 6(b)] as [51]
ing is done electrically and not mechanically. An array of

s (t) = '
such plasma antennas may be equipped with smart signal A cos (2rft + i) for x
. (16)
processing algorithms [50] so that electronic RF beam steer- 0 for PRT - x
ing can be achieved. The refractive index for uniform plasma
is given by n = 1 + ~ 2p / (~ ( jv - ~)), where ~ p is the plasma Similarly, the frequency can also be modulated with
frequency, and v is the electron-neutral collision frequency. the help of a PRT signal. The different schemes of frequen-
cy modulations are pulse-frequency modulation, linear-
RADAR WAVEFORM DESIGN frequency-modulated continuous-wave (LFMCW), and
AND ITS PROCESSING TECHNIQUES stepped frequency modulated CW (SFCW or SFMCW).
Radar waveform design more suitable for accurate remote Pulse-frequency modulation is applicable in pulsed-radar
sensing applications, and various radar signal processing systems, while LFMCW and SFMCW are for CW radar
techniques available for precise detection, imaging, and systems. In pulse-frequency modulation (also known as
classification of the targets, are reported in this section. a pulsed-Doppler signal), the frequency of the RF signal is
linearly increased/modulated [see Figure 6(c)] within the
WAVEFORM DESIGN period of x width so as to get the increased range resolu-
The radar waveform used in early-stage remote sensing radar tion (equal to C/2b, where b is the modulation band-
systems is a CW RF signal as s (t) = A cos (2rfc t + i), which width) without reducing the ­actual x as
is continuously transmitted over all the time, as shown in Fig-

s (t) = (
ure 6(a). Due to some limitations of CW signal transmission, A cos (~ 0 t + rbt 2) ~ r for x
(17)
the modulation of it was introduced. Later, in remote sensing 0 for PRT - x

1 1 1
0.5
Amplitude

0.5 0.5
Amplitude

Amplitude

0 0 0
–0.5 –0.5 –0.5
–1 –1 –1
0 200 400 600 800 0 200 400 600 800 0 200 400 600 800
Time Time Time
(a) (b) (c)
1 1 1
0.5
Amplitude

0.5 0.5
Amplitude

Amplitude

0 0 0
–0.5 –0.5 –0.5
–1 –1 –1
0 200 400 600 800 0 200 400 600 800 0 200 400 600 800
Time Time Time
(d) (e) (f)
1 1
AF of P4 Code

Amplitude

0.5 1
Amplitude

0.5
0 0.5
0
–0.5 –5
–1 20 10 0 –10 0 ncy –0.5
0 200 400 600 800 –20 5 reque –6 –4 –2 0 2 4 6
Delay F Normalized Frequency
Time
(g) (h) (i)

FIGURE 6. Remote sensing radar waveforms. (a) CW, (b) an amplitude shift-keying wave, (c) a pulsed Doppler wave, (d) an ICW,
(e) SFMCW, (f) a Barker-coded wave, (g) Frank-coded wave, (h) polyphase-coded wave, and (i) an orthogonal frequency-division
multiplexing (OFDM) wave.

76 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 where b is the chirp rate, ~ 0 is the starting frequency of the into N equal parts. This gives a Frank code compression ra-
chirp, ~ r is the end frequency, and t are the time samples tio equal to p = N 2 [55]. Polyphase codes P1, P2, P3, and
within the x. The frequency chirps used on the LFMCW P4 are another advanced technique used to generate the
signal are series of periodic linear signals, linearly modulat- phased-coded remote sensing radar waveform, and their
ing its frequency content during every PRT period [see phase values can be given by
Figure 6(d)], and the same can be generated as [52]
-r
i i M, j M = M [M 1 - (2j M - 1)] [(j M - 1) M 1 + (i M - 1)]
1
s (t) = A cos (~ 0 t + rbt 2) ~ r (18) -r
i i M, j M = 2M [2i M - 1 - M 2][2j M - 1 - M 2]
2
r 
where t are the time samples within the PRT. The down i i N = N (i N - 1) 2
s
chirp LFMCW, or triangular (up-down chirp) LFMCW r (i N - 1) 2
signals, are also applied in various remote sensing radars. i iN = Ns - r (i N - 1) (22)
For an SFCW wave, the frequency of the signal is increased
with a step value for a specified time period [see Figure 6(e)] where i M = 1, 2, ..., M, j M = 1, 2, ..., M, M 1 = 1, 2, 3, ...,
within the PRT as M 2 = 2, 4, 6, ..., i N = 1, 2, ..., N, and N s is the compres-
sion ratio. An AF generated for the P4 code is shown in Fig-
fn = f0 + nDf (19) ure 6(h). A fourth generation of waveform design is based
on the combined modulation of the frequency and phase
where fn stands for the frequency of the nth pulse, f0 information of the radar signal, which is known as the
is the starting carrier frequency, and Df is the fixed- orthogonal frequency-division multiplexing (OFDM) modu-
frequency increment. These step waves are frequency lation scheme. OFDM is an up-and-coming modulation
modulated/incremented to get a final SFMCW waveform. scheme in which the modulation is achieved using mul-
SFCW waveform-based remote sensing radars have sev- tiple frequency components. A waveform generated for
eral advantages, like better resolution for closely spaced OFDM is shown in Figure 6(i). OFDM gives a high pulse
targets, higher ranging accuracy, and significant reduc- compression rate by dividing the entire pulsewidth into
tion of clutter effects [53]. multiple narrow widths, transmitting the modulated RF
The third generation of radar waveform design are phase signal in parallel as [54]
modulation techniques, which include three types: the
N1
Barker, Frank, and polyphase (P1, P2, P3 and P4) codes.
These phase modulation schemes are also suitable for
s n (t) = / S n,k e j2rkDft, 0 # t # Ts (23)
k=0
pulsed-radar systems. The Barker code (B 1 through B 13) is
used for binary phase coding as where Ts is the symbol duration, Df is the subchannel space,
and N is the number of subchannels of the OFDM signal.

s (t) = (
A cos (2rft + i) for “1”/“ + ”
(20)
- A cos (2rft + i) for “0”/“–” AF
An AF (also known as an uncertainty function) is an im-
where the “1” or “+” and “0” or “–” denotes the elements portant phenomenon to have an effective analysis of
of the Barker sequence. This phase coding divides the designed radar waveforms in the 2D plane: delay, i.e.,
pulse of the xl width into smaller pulses of Dx width as range, and Doppler, i.e., velocity. The results of the AF are
Dx = xl /N, where N is the length of the Barker code. Then, equal to the absolute value of the envelope of matched
zero radians or r radians is the phase of each subpulse filter outputs. This function is important for analyzing
chosen. Binary phase codes have a compression ratio given the accuracy/resolution of both the range and velocity of
by p = xl /Dx [55]. The other phase modulation method is moving targets. The AF is defined by the properties of the
the Frank code, which follows the phase modulation with pulse of the filter and not any particular target types/sce-
more than two phase values. The modulation phase values nario. For a transmitted signal s (t), the AF is given by [58]
are based on fundamental phase increments, known har-
monically as
| X (x, fd)| = #- +33 s (t) s ) (t + x) exp ( j2rfd t) dt (24)
2r
W a, b = M (a - 1)(b - 1)(21)
where x is the delay, and fd is the frequency shift. The AF
where M = L is the dimension of the code matrix, L is the analysis has to be made for any newly designed radar wave-
number of code elements, a = 0, 1, ..., M is the row number form before transmitting it for the purpose of remote sens-
of the code matrix, and b = 0, 1, ..., M is the column number ing. The AF represents the distortion of a returned pulse due
of the code matrix. In the Frank code-based phase modu- to the receiver-matched filter (commonly but not exclusive-
lation, a single pulse having pulsewidth xl is divided into ly used in pulse compression radar) of the return from a
N subpulses, after which each subpulse is again divided moving target.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 77

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 PULSE COMPRESSION AND STRETCH PROCESSING N times in the N-pulse CPI technique, while in the case
In addition to radar waveform design/generation tech- of N-pulse NCPI, the SNR is increased by N times. After
niques and their delay-Doppler domain AF analysis men- improving the SNR, the signal is applied for further pro-
tioned earlier, pulse compression techniques are also cessing for T-F analysis. The Fourier transform is usually
important at the receiver chain, which are normally per- applied to the stationary signals to transform them from
formed by a matched filter and stretch processing. Matched time domain to frequency domain. If the signal is not sta-
filtering is the correlation of a known delayed signal with tionary and evolves with time, then a local frequency pa-
another unknown signal, and it is used for signal detection rameter is used to look at a small part of the signal under
in radar systems. For a signal s (t), the impulse response for the assumption that the part of the signal under consider-
its matched filter is given by [56] ation is stationary. A popular technique used for T-F analy-
sis, i.e., F (~, x), of the signal f(t) is the short-time Fourier
h (x) = ks (D - x)(25) transform (STFT) as [59]

where k and x are arbitrary constants. The matched filter F (~, x) = #- 33 f (t) } ) (t - x) e -j~t dt.(28)
is the first technique applied to significantly improve the
signal-to-noise ratio (SNR) as The Gabor transform is a special case of STFT where
the window } (t) is a Gaussian function. Modulation and
aN
Sk
= a N k 2B IF x (26)
S
transformation of the window function } (t) gives us the
out in
basis function for this transform, and the Gabor transform
where (S/N) out is the output SNR and (S/N) in is the input is given by [59]
SNR, B IF is the bandwidth, and x is the pulse duration.
Another pulse compression technique is stretch process- F (~, x) = #- 33 e - r(t - x) e -j~t dt.(29)
2

ing [56], which uses linear frequency modulation along

with low-bandwidth signal processing techniques to The wavelet transform is another such transform tech-
achieve a high-resolution range. A local oscillator (LO) nique that converts the signal from time domain to frequency
linear frequency-modulated signal is then multiplied to domain and vice versa. There are several subtypes of wavelet
stretch the received signal. Both signals should have the transforms, and to take the wavelet transform of a signal we
same sweep rate. The output signal is given by summing need to have a mother wavelet, which is given by [59]
up all the scatters where every single scatter is the sum of
} a, b (t) = c m } a a k . (30)
1 t-b
the received signal and the LO signal, and the output sig-
a
nal of a stretch processor is given by [57]
With this as the mother transform, the wavelet trans-
P
x [m] = / A k v k exp ( j2r ( fc 2Rc k + M
B 2R k
c i [m]
form of a signal is given by
k=1

+
2Vk Ts fc
c i [m] +
B 2Vk Ts
M c i [m ] 2
)) + w [m] (27) W } f (a, b) =
1 #- 33 f (t) } * a t -a b kdt (31)
a
where m = 1, 2, .., M, M is the number of samples in the where a and b are the scaling and translation parame-
pulsewidth, i [m] = m - (M/2), B is the bandwidth of the ters, respectively, and, if they are continuous, the wave-
linear frequency-modulated signal, Ts is the sampling time, let transform will also be continuous. Another popular
R k is the distance from the kth scatter to the scene center, and more accurate method for T-F analysis is the Wign-
c is the speed of light, Vk is the radial velocity of the kth er–Ville distribution (W VD), which was first introduced
scatter toward the radar, v k is the RCS of the kth scatter, by Eugene Wigner in 1932 for quantum thermodynam-
fc is the carrier frequency, and w is the noise vector. Apart ics and reintroduced by Jean-André Ville in 1948 for a
from these pulse compression techniques, other techniques signal’s T-F analysis. The W VD provides a good tempo-
based on an inverse fast Fourier transform (IFFT) could ral resolution at HF and good frequency resolution at
also be used for pulse compression in the case of frequency low frequencies as it has variable resolution over the T-F
radars, e.g., SFMCW GPR systems, in which the frequency plane as [60]
response of the subsurface targets are initially collected,
and then the time-domain impulse is constructed using the WVD X (t, f ) = #- 33 X` t + 2x jX ) ` t - 2x je -j2rfx dx (32)
IFFT pulse compression technique.
where x is the time-delay variable, and X (t) is the complex
T-F ANALYSIS signal associated with x (t), with the complex part obtained
Followed by the matched filter-based SNR improvement, by the Hilbert transform. A variation of the WVD, called
the subsequent SNR improvement techniques are coherent smoothed pseudo WVD, is often used as it restrains cross-
pulse integration (CPI) and non-CPI (NCPI), and both are term interference and enhances energy distribution con-
known as processing gain techniques. The SNR is increased centration as [60]

78 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 SPWVD g, h, X (t, f ) also becomes complex. Several radars, like SAR are ca-
= #- 3 #- 3 X` t + 2x jX * ` t - 2x jg (v) h (x) e -j2rfx dvdx
3 3 
(33)
pable of 3D imaging. Simple 2D image processing of
radar is done by combining the range and Doppler pro-
where g (v) is the time-smoothing window, and h (x) is the file analysis. Several factors, like rotation rate, i.e., the
frequency-smoothing window. Both should be real sym- distance of the target from its rotation center, and the
metric functions with g (0) = h (0) [61]. wavelength of the transmitted signal, are considered
while processing the data to obtain a 2D image of the
RANGE-VELOCITY MAPPING target [64]. However, radars that do 3D imaging use
Range-velocity (RV) mapping can be constructed with sum-millimeter waves (mm-waves) to take advantage of
range and velocity information obtained by processing the high spatial resolution [65]. 3D radars are mostly pre-
time interval and frequency difference between the trans- ferred to form multiple beams to take care of multiple
mitted and received signals [62]. The FMCW waveform of different missions/sensing/targets simultaneously. 2D
radar systems gives a straightforward approach to detect SAR, ISAR, and 3D RF sensors have found a wide variety
both the range and velocity profiles of targets. The trans- of applications, from military activities to agricultural
mitted FMCW wave can be expressed as [63] management. Hence, today, 2D and 3D image process-
ing of SAR have become essential activities. In 2D and
B
f (t) = fc + T t (34) 3D target imaging, before proceeding into core signal
processing, several preprocessing techniques, like noise
where fc is starting frequency, B is the bandwidth, and T is filtering, median filtering, Gaussian filtering, range mi-
the time duration of frequency modulation. This transmit- gration, clutter removal, SCR/SNR improvement, and so
ted signal scattered from the targets is mixed with the local on, have to be carried out [66]. This is followed by sub-
reference chirp signal to get the IF, i.e., FIF, which directly sequent image processing algorithms to extract the main
gives the range (R) as information from the constructed 2D/3D target images.
The emerging method for doing this is using deep learn-
FIF cTP
R= (35) ing (DL) algorithms and generative networks. Therefore,
2b
the first step is data augmentation, where the data are
where R is the range, FS is the sampling frequency, c is the generated by analyzing various parameters of the signal
velocity of propagation of EM waves, TP is the PRT, and b received, which carries the targets’ structural data and
is the chirp rate. Taking the FFT along the PRT axis, known demands the next step to translate the data to an image
as the first FFT, gives the range information. In FMCW ra- form [66].
dar, the phase difference Dz between the transmitted and
received signals keeps appearing for the moving target with DIRECTION OF ARRIVAL FINDING
respect to its velocity; otherwise, for the static target the Dz One of the critical branches of radar signal processing is the
is constant, which can be given by [63] estimation of direction of arrival (DoA). It involves find-
ing the direction in which the signal is traveling. In several
Dz = 2r ; c + c c + Tp m tE (36)
2fc R 2fc v 2BR
remote sensing applications, passive and active RF sensors
are deployed to find the DoA of the incoming EM signal.
where v is the velocity of the target. Thus, every received Generally, an antenna array is used for the DoA systems
signal changes its phase, which produces a frequency that ultimately process phase differences and compute
equal to the DoA details. Different algorithms are used to find the
DoA, including popular algorithms such as beam scan,
2fc v 2BR
f= c + Tc . (37) minimum-variance distortion less response, multiple sig-
nal classification (MUSIC), estimation of signal parameters
By solving this equation taking the FFT along the sample using rotation invariance techniques (ESPRIT), and eigen-
axis, known as the second FFT, we can obtain the velocity value vector clustering [67]. The angle of DoA of EM waves
information of the targets. Sketching the targets’ range and in terms of antenna elements, antenna gaps, and phase
velocity information obtained as detailed in (35) and (36) shifters can be given by
on a 2D plane gives the individual RV image/signatures of d sin (i) 2d sin (i) (N - 1) d sin (i)
all targets [63]. a (i) = [1, e i2r m , e i2r m ,..., e i2r m
H
] (38)

2D/3D TARGET IMAGING where a is the steering vector representing phase shifts
In general, a radar display is a 2D plane, either in range between antennas in the array, d is the distance between
and cross range, or in range and angled, or in range and two antennas, H is the Hermitian matrix, and N is the
velocity. In today’s advanced radars of remote sensing, number of antennas. MUSIC is a popular subspace-
the display is in a 3D plane as in range, cross range and based algorithm that was proposed in 1986. It provides
angled/coordinate, and the associated signal processing an estimate of the number of signals, polarizations,

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 79

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 N
DoA, strengths, cross correlations of directional wave-
H (~) = < / G (R i, X k ; ~) t k (~) G (X k, R j ; ~)F
L
(42)
forms, and strengths of noise, interference, and so forth. k=1 i, j = 1
ESPRIT is also a subspace-based algorithm. Proposed in
1989, it is an improvement over the MUSIC algorithm. where G is the Green function, R j denotes the antenna’s
ESPRIT uses rotational invariance among signal sub- center space points, and X k denotes target-centered space
spaces, which demand lesser computational and storage points. Today, the time-reversal signal processing technique
requirements [67]. is popular because it 1) creates an optimal carrier signal for
remote sensing, 2) reconstructs a source event, and 3) fo-
STAP cuses high-energy waves to a point in space. Thus, the time-
Simultaneous processing of antenna-array-received sig- reversal technique has become more general in various
nals by multiple phase-coherent waveforms is known as remote sensing applications, including time-reversal trans-
STAP. Targets that are hidden by main lobe or side-lobe mission in radar, the time-reversal detection technique,
clutter or missed due to jamming environments can be and time-reversal MIMO radar.
detected using the STAP technique [68]. For an array on
N antennas with a PRF of M, an M # N snapshot can be ARTIFICIAL INTELLIGENCE/COGNITIVE ALGORITHMS
formed covering the target, and it can be splinted into FOR SIGNAL PROCESSING
components having the target and other components In today’s complicated targets scenarios, particularly swarms
having interference/clutter and noise, which makes STAP of sUAV/drone targets, processing, detecting, and classifying
a unique radar signal processing algorithm, especially in such targets using classical signal processing techniques is
airborne radars of high-clutter scenarios. If v is a space– almost impossible. Thus, the necessity of detecting/classify-
time steering vector, the optimum processor, i.e., the ing them invites the involvement of machine learning (ML)
weight vector, is given by [68] and DL techniques. Artificial intelligence (AI)/cognitive al-
gorithms have been put to use in modern radars for target de-
w = R -1 v (39)
tection and feature identification/classifications. Initially, a
where R is the covariance matrix of the interference, which radar image is obtained and processed, with some enhance-
includes jamming, clutter, and the noise component of ra- ments and removal of noises. Then, AI/cognitive-based ap-
dar-received data. The optimum processing weight vector of proaches like decision-tree-based target detector, feature
[NMx1] in an STAP algorithm can be given by parameter or vector-based target detector, and machine
learning classifier are implemented. The orientation, fre-
w = :# s c (z) v c (z) v Hc (z) dz + I M 7 M j + v 2 I MND v c (40)
3 -1
quency of transmitted signal, or processing techniques can
-3
be changed by the AI/cognitive algorithms for accuracy-
where w is the weight vector, v c is the space–time steering increased target detection and classification [71]. The for-
vector, H is the Hermitian matrix, s c is clutter power spec- mulation of a deep convolutional neural network (DCNN)
tral density, v is the receiver noise power per element per model, DIATnRadSATNet, is given by [72]
pulse, and M j is the jammer spatial covariance matrix.
DIAT-RadSATNet (X) = b (m FC (m d (m g (m RsE k
.
TIME-REVERSAL TECHNIQUE (m M (m b (v (m c1 (X ) W i + B i))))))))) (43)
Time reversal is a signal processing technique used in radars
for active remote sensing. Time-reversal imaging uses N an- Paths one, two, and three of the residual squeeze and ex-
tennas to sequentially probe one or more than one targets. pand (RsE) block in DIATnRadSATNet can be expressed by
Backscattered returns at all antenna locations yield the fre-
quency response matrix H (~) as m p1 (X) = v (m cj + 3 (m X ) W i + B i))(44)
m p2 (X) = m b (v (m cj + 1 (v (m cj (m X ) W i + B i)))))(45)
H (~) = {H ij (~); 1 # i, j # N}(41)
m p3 (X) = m b (v (m cj + 2 (v (m cj (m X ) W i + B i))))). (46)
where H ij (~) is the response between the ith element and
jth element of the array [69]. In MIMO radar, these time- Finally, the RsE layer can be formulated by combining
reversal techniques can be used to automatically match the (42)–(44) appropriately as
waveforms to a scattering channel, further improving the
performance of radar detection [70]. These time-reversal m R s E k (X) = (m p2 (X) 5 m p3 (X)) 5 m p1 (X)

processes are used for effective carrier signal generation and i ! {1, 2, ..., n}; j ! {2, 6, 10, 14}; k ! {1, 2, 3, 4} (47)
focus high-energy waves to a point in space. The frequency
response matrix H (~)(~) as a function of geometry and where B i is the bias matrix, W i is the weight matrix, X is the
physics of wave prorogation between antenna elements input image, m cj is the convolution layer for the j layer in-
and the scatterers can be expressed in association with the dex, v is the rectified linear unit activation, m b is the batch
Green function as normalization layer, m M is the maximum (max) pooling

80 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 layer, m RsEk is the RsE, m g is the global average pooling lay- REAL-TIME SIGNAL PROCESSING
er, m d is the dropout layer, m FC is the classification layer, b With radars being used heavily for remote applications
is the softmax activation, m X is the output feature map of for multiple applications all around us, processing of the
the previous layer, and m p1 (X), m p2 (X), m p3 (X) are the out- data is a crucial activity when we do it in software plat-
put tensors [73]. Cognitive radar or cognitive-based radar forms due to serial executions. Storing, transferring, and
signal processing has emerged in the past 15 years, con- processing the data becomes quite laborious. Real-time
taining the synergy of cybernetics, waveform diversity, and radar signal processing for high-speed remote sensing
knowledge-aided signal processing as a new vision for fu- applications invite hardware-based embedded solutions.
ture radar systems and remote sensing applications while Due to advanced developments in the field of very large-
addressing today’s requirements of performing remote scale integration technology, software-driven embedded
sensing in a dense RF noise/spectrum distributed environ- platforms, RF system on chip (RF-SoC), network on chip
ment [74]. This cognitive radar is a new paradigm that con- (NoC), and many more, hardware-embedded platforms
stantly observes the surrounding environment, inferring are now available, through which real-time signal pro-
properties about the environment based on observations, cessing becomes realizable/achievable. Hence, several
and modifies its observation of the environment through radar applications have onboard data processing hard-
changing waveforms and antenna-pointing directions, ware platforms, like CPUs, GPUs, RF-SoCs, and field-
maintaining real-time hypotheses about the environment programmable gate arrays (FPGAs). These hardware
to adaptively control the form and parameters of the actual can achieve extremely high processing rates, making
sensing procedure. the data processing activity more efficient and real time
In a tracking problem, the kinematic characteristics of [75]. In addition to the huge digital resources, intellec-
the target, and the model of a physical system in terms of tual property cores, and digital signal processing design
input, output, and state variables, can be given by blocks available in today’s embedded platforms, FPGAs
come with the possibility for implementation of pipe-
x k + 1 = f (x k) + w k (48) lined, parallel architectures that reduce the latency and
z k + 1 = h (x k + 1) + n k . (49) significantly increase overall throughput. The software-
defined radio (SDR) platform comes with on-the-fly, re-
This optimum solution upon solving (48) and (49), configurable baseband internal modules/resources, such
along with the Gaussian model and observation noise, is as analog-to-digital converters (ADCs), digital-to-analog
converters (DACs), RF up mixers, power amplifiers, low-
xt k + 1/k = Fxt k/k (50) noise amplifiers (LNAs), and so forth, which can be
Pk + 1/k = FPk/k F T + Q K (51) driven by software that supports the generation of high-
phase, stable remote sensing signals and the carrying out
xt k/k = xt k/k - 1 + K k (z k - Hxt k/k - 1)(52) of real-time radar signal processing. The Jetson Nano em-
Pk/k = Pk/k - 1 - K k S k K KT (53) bedded platform and ultra large-scale integration FPGA
boards are now available for the real-time implementa-
where tion of AI architectures [76].

S k = HPk/k - 1 H T + R (54) DEVELOPMENTS IN RADAR SYSTEMS’

T -1
K k = Pk/k - 1 H S (55)
K
ENGINEERING DESIGN
Radar and remote sensing technologies have evolved through
where x k is the target state at the kth time instant; z k is the several generations of transformations for their RF front-end
radar observations at the kth time instant; w k and n k are engineering designs. A review on them covering technol-
the model and observation noises, respectively; xt k/k is the ogy revolutions corresponding to vacuum tubes, wave-
target-state estimation; xt k + 1/k is the target-state prediction; guide devices, solid-state devices (SSDs), RF-SoC/RF-NoC
Pk/k is the covariance matrix of the target-state estimation; integrations, photonics microwave devices, and quantum
and Pk + 1/k is the covariance matrix of the target-state pre- microwave devices are reported in this section.
diction. F are the Jacobians of the system equation, and H is
the measurement equation transfer function. In this prob- VACUUM TUBE MICROWAVE DEVICES
lem, the covariance matrix of the target-state estimation In the beginning of remote sensing technology, the earli-
Pk/k is cognitively updated every iteration in accordance est radar system engineering belonged to the generation
with the real tracking states. Therefore, in cognitive-based of vacuum tubes. These were vacuum glass tubes that
remote sensing, the prior and current knowledge are quan- produced X-rays from cathode rays striking a metal target
tified and updated in real time to optimize the sensing [77]. The most notable attempt was the Coolidge tube, de-
procedure in a closed-loop operation, thereby optimizing veloped in 1913. It had a heating tungsten filament, which
radar performance and making efficient use of finite sens- produced electrons from thermionic effect. These elec-
ing resources. trons travel from cathode to anode due to high potential

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 81

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 difference and hit the anode to produce X-rays [77]. The almost all the operations of radar subsystems, such as
next major milestone in this field was the klystron [see power dividers, frequency meters, microwave couplers,
Figure 7(a)], developed in 1937. It works on the principle isolators, circulators, microwave junctions, variable at-
of velocity modulation. In a klystron, electrons are accel- tenuators, tunable probes, slotted sections, termina-
erated from the cathode by potential difference and, as tors, and so on. Those waveguides are of different shape
they travel, the electrons experience velocity modulation (rectangular, parallel plate, and circular) and size based
by a sinusoidal RF signal. This velocity-modulated beam on the frequency of operation. They have several advan-
creates an output cavity at the load, which slows down tages, like high power-handling capacity and low losses,
the beam. The kinetic energy of electrons during this pro- but they are bulky and expensive. Those waveguides
cess is transmitted as EM waves [78], which created paths follow different modes, like transverse electric (TE) or
for remote sensing. Magnetron was the next generation transverse magnetic modes, which are characterized by
of vacuum-tube-based RF generators, which also used the presence of TE or magnetic field components. During
resonant cavities under EM fields to generate microwave WWII, radars were built with several such waveguides,
waves. Developments happened in vacuum tube genera- where the power dividers and directional couplers divide
tions in the 1940s, highly supported to make the earlier or combine power from one to multiple ports. Based on
radars, and had an significant impact on WWII [79]. After the orientation of the rectangle of the waveguide and its
the period of vacuum-tube-based X-ray generators came connection with T junctions, it is one of two types: E-
the era of semiconductor-based EM wave generators. One plane or H-plane waveguide. The necessity of inventing
of the earliest and most notable attempts was the Gunn other technologies due to waveguide devices’ bulky size
diode, developed in 1962. Gunn found that for III-V semi- to design all the microwave components created a path
conductors, there was a region of negative effective resis- to solid-state technology [81].
tance, where the electrons oscillated and produced EM
waves [80], which enabled research in the field of semi- SOLID-STATE MICROWAVE DEVICES
conductor technology for developing the microwave com- Starting with the Gunn diode, radars were made using
ponents to make remote sensing radars. a semiconductor, and radars based on SSDs came into
existence. These SSDs have several RF components, like
WAVEGUIDE DEVICES frequency-tunable voltage-controlled oscillators (VCOs),
Followed by vacuum tube and semiconductor technolo- diode RF mixers, power dividers, in-phase/quadrature-
gies, waveguide-based microwave components/devices phase (IQ) signal generators, LNAs, hybrid couplers,
structures that allow waves to travel mainly in one di- transmit–receive switches, power amplifiers, and so
rection were developed. Almost all of these microwave forth, as shown in Figure 7(c). All these SSD microwave
devices/components are waveguide based and passive, devices are smaller in size and can be operated with low-
other than the source and detectors, as shown in Fig- power sources. A VCO is an SSD oscillator that produces
ure 7(b). The waveguide components were developed for an output signal whose RF signal is controlled by an

(a) (b) (c)

(d) (e) (f)

FIGURE 7. Developments of radar subsystems/components engineering. (a) Vacuum tube technology, (b) waveguide technology, (c) SSD
technology, (d) RF-SoC technology, (e) photonic microwave technology, and (f) quantum microwave technology.

82 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 external input dc voltage. An RF mixer is a nonlinear or with less susceptibility to noise [86]. The RF-NoC is a
time-varying component, which is a multiport network framework for controlling a high-dense different net-
that generates the output by adding or subtracting the work routing protocols as signal applications running
frequencies of two input (LO and RF) signals. LNAs are in a single chip [87].
amplifiers that do not introduce significant noise while
increasing the power of the radar’s weak echo signal. PHOTONICS MICROWAVE TECHNOLOGY
The other type of active RF component used in the radar One of the most modern technologies under recent devel-
is a power amplifier, which amplifies the RF signal prior opment is RF/microwave photonics, which is similar to a
to transmission by a suitable antenna. SSD RF compo- conventional electrical technology but mostly with insu-
nents give relatively minimum heat dissipation, even lators (fiber-optic cables and lights) instead of conducting
while operating for high-power applications. A power mediums (RF coaxial cables and currents). As the carrier is
divider is another passive component that splits an in- the optical light and the medium is the insulators, photon-
put signal into multiple signals with low split losses. ics-based systems are completely free from the effects of
An attenuator is supposed to handle the opposite func- EM interference, and the photonics devices are very small
tion of an amplifier as it limits the strength of a sig- in size. All the photonics devices are capable of operating
nal [82]. A digital direct synthesizer (DDS) is another the extremely high instantaneous bandwidth, e.g., the in-
very successful device under the SSD technology and stantaneous RF bandwidth of a Mach–Zehnder modulator
synthesizes the radar waveform based on digital tech- is 40 GHz. Due to these unique advantages, RF/microwave
niques. They provide high phase-stable RF signals, wide photonics replaces all the electronic subsystems/compo-
frequency ranges, and fast settling times. The six major nents. Thus, RF photonics is an interdisciplinary field
waveforms of DDSs are pulse output, fractional divider, that combines microwave and photonic technologies [see
sine output, triangle output, phase interpolation, and Figure 7(e)] for the generation, transmission, processing,
jitter-injection DDS. During the same era, microstrip and control of RF signals ultimately taking advantage of
antenna design technology was developed/adapted. All the broad bandwidth, high accuracy in frequency tuning,
these SSD components are configured using frequency- high-frequency range, low coupling/connection loss, and
limited RF coaxial cables with different types of con- negligible short-range transportation loss [84]. In the last
nectors of different impedances, mostly 50 X. Thus, few decades, numerous solutions have been proposed and
the remote sensing radars developed since this era are demonstrated, such as microwave photonics components,
lightweight and compact, which supported building photonic-assisted microwave high-frequency/bandwidth
high-frequency (X band and above) smaller-size radars; exciters, photonic-assisted RF signal processing, photonic
lightweight, vehicle-mounted ground-based radars; aer- ADCs, photonics-based frequency/phase/amplitude mea-
ial platform-carrying radars; satellite platform radars; surements, photonics-based DoA measurements, and true
and so on [83]. time-delay-based RF beamforming. The realization of
RF photonic systems based on photonic ICs (P-ICs) has
RF-SOC/NOC INTEGRATED CIRCUIT TECHNOLOGY also been recently demonstrated with clear-cut photonics
Several advancements have happened in SSD technol- circuits design, circuits fabrication, and implementation
ogy, and today it is possible to design/fabricate devices approaches. Application-specific P-ICs for RF photonics
of subnanometer size. All these advancements enabled applications and programmable integrated microwave
the possibility for high-dense integration of program- photonics components, RF-optical modulators, optoelec-
mable RF integrated circuits (ICs), which have given a tronic RF generators, and tunable optical filters have also
paradigm transformation in remote sensing radar tech- been demonstrated [85].
nology. Then, the radar system is known as SDRs, in
which all the subsystems that are fabricated in a single QUANTUM TOWARD MICROWAVES
RF chip are programmable and on-the-fly reconfigu- Another emerging remote sensing technology is the cross
rable. This integrated semiconductor-based technology research between microwave photonics and quantum tech-
includes RF-SoC and RF-NoC with high-density inter- nology, which is known as quantum microwave photonics, as
nal resources. An RF-SoC is the RF analog and digital shown in Figure 7(f). Quantum toward microwave technol-
circuits that are integrated within a single chip with a ogy appears that it can achieve features or functions that
provision for implementing the programmable logics are very complex or even not possible using traditional
(design flexibility) and processors (high-speed compu- microwave and/or even RF photonic technologies. The ba-
tations), as shown in Figure 7(d). VCOs, LNAs, pulse sic operation of quantum microwave block building, i.e.,
modulators, frequency multipliers, ADCs, DACs, and microwave quantum single-photon generation, and detec-
other similar components are fabricated on the chip, tion schemes is a current theoretical/experimental achieve-
making the radar system smaller in size, easy to pro- ment in the realization of quantum microwave systems.
gram, at good phase stability, at a high sampling rate, Research has been triggered in the quantum microwave
at instantaneous high bandwidth, and so forth, and field for the demonstration of weak microwave signal

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 83

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 detection, microwave photonic phase shifting/filtering, require very low-range radar, which would sense a few cen-
nonlocal time-frequency mapping, RF compressed sens- timeters. Today’s automotive radar systems are capable of
ing and quantum key distribution, quantum microwave detecting targets, i.e., the vehicles around them, at different
sensing, quantum microwave communication, microwave- ranges, and they use two frequency bands: one around 24–
quantum information processing, and so on. Impressive 29 GHz for shorter ranges and the other around 76–81 GHz
progress has happened, both in quantum microwave and for longer ranges [89], [90]. Figure 8 shows the schematic
superconducting quantum circuits, providing a platform diagram of an automotive radar, which is driven by a VCO
for manipulating microwave photons in a controlled fash- with a PRF of a few microseconds. The received signal is
ion. Emerging microwave quantum technology will soon amplified by the LNA and passed into the signal processor
deal with microwave signals in a completely different way, through the mixer [91]. In an automotive radar, the range
which will make paradigm changes in numerous applica- and relative velocity of the vehicle can be calculated by [92]
tions such as broadband wireless access networks, sensor
networks, radar, satellite communications, remote sensing, cDR (N - 1) Dz - rl
R= r $ (56)
and warfare systems. c - 4 (N - 1) fshift DR
(N - 1) Dv cDz - 4rfshift DRl
v= r $  (57)
CONTEMPORARY SOCIETAL APPLICATIONS c - 4 (N - 1) fshift DR
OF RADAR SYSTEMS
This section explores the direct impact of remote sensing radar where c is the speed of light, DR is the range resolution,
technology on societal applications. Dv is the velocity resolution, N is the number of frequency
steps in each transmit signal sequence, Dz is the phase dif-
AUTOMOTIVE RADAR ference between the two transmit signal sequences, l is the
Safety has always been a concern for automobiles, and ra- index of the spectral peak detected in the Fourier spectrum,
dars have proven to be an effective solution to ensure it. and fshift is the frequency shift between the two transmit
Research on automotive radars began in the 1960s and sequences. Further, an autonomous self-driving vehicle is
today’s automotive radar technology converts the vision a concept that depends heavily on automotive radars. To
of fully autonomous cars into a reality [88]. Today, many drive the car autonomously, radars have to detect the dis-
cars have ADASs, which have several subsystems perform- tance and direction as well as the velocity of the cars and
ing specific functions/tasks that are mainly driven by the other obstacles around it. Hence, FMCW MIMO radars
radars like adaptive cruise control, automatic emergency are used for this application [93], [94]. MIMO radars are
braking, lane change assist, and so forth [89]. Adaptive broadly classified into two types: separated MIMO radars
cruise control demands short-range radars, i.e., up to a and co-located MIMO radars. The first ones take more space
few hundred meters. Subsystems like lane change assist due to their bigger antenna size to sense different targets/
and blind spot detection require medium range, which is objects of varying RCS, while the second ones are compact
of a few hundred meters, while subsystems like park assist in size and hence preferred, even for small cars.

BIOMEDICAL RADAR
Remote sensing radar finds several
applications in the biomedical indus-
Power try, such as human vital sign moni-
Tx Switch
Amplifier Transmit toring, fall detection, vision sensors,
Antenna brain strokes, tumor detection, con-
tactless human body scanning, hu-
Power man body 3D imaging, complete
VCO Receive
Splitter human body function detection/­
Antenna
measurements, and so forth. Bio-
medical applications usually require
short-range and low-power Doppler
Rx Switch Mixer LNA
or FMCW radars [21], [96], [97]. The
combination/hybrid of Doppler and
FMCW modes in a single waveform
allows for taking advantage of both
Signal
PRF Processing waveform schemes in a single hard-
ware setup and gives accurate re-
sults in the measurements [21]. Vital
FIGURE 8. The top-level design layout of an automotive radar. Tx: transmitter; Rx: receiver; signs, as the name suggests, are basic
VCO: voltage-controlled oscillator. functions, like heart or respiration

84 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 rate, which indicate the state of the body. The range and populated area to measure/predict various parameters per-
RCS of the radar used for vital signal monitoring can be ob- taining to climate change [103]. Weather radars use fre-
tained from the signal-received echo. The respiration rate of quencies ranging approximately from 3 to 10 GHz [104]. It
the average human body is usually expected to be approxi- is important to choose the right band of frequency based on
mately 10–12 breaths per minute (bpm), and the heart rate the required signal attenuation, range, and indented appli-
is expected to be roughly 60–100 bpm, which can be accu- cation [105]. S-band (2.7–2.9 GHz) weather radar provides
rately measured by using a biomedical radar without plac- a range of roughly 200 km, and this RF band is less affected
ing any electrodes on the human body [98]. The Doppler by attenuation. The C band (approximately 5.6 GHz) can
and FMCW radar systems, with frequencies ranging from detect rain up to a range of 100–150 km, and it undergoes
hundreds of megahertz to mm-wave, are used to measure a significant attenuation when compared to the S band.
vital signs. However, a 228-GHz signal was also used for On the other hand, the X band has shorter ranges, roughly
the same purpose as it has much shorter wavelength ad- 50 km, and stronger attenuation due to rain compared to
vantages [96]. Sensing smaller displacements in the human the S and C bands. X-band weather radar has the lowest cost
body is easier with shorter wavelengths. The RCS improves and most sensitive radar. Polarization is an important fac-
with higher frequency as the HF beam is collimated to the tor that should be considered, particularly in weather radar
point/region of interest on the human body for measure- as it deals with the direction of the transverse movement of
ment [96]. Biomedical radar normally uses different radio the constituent fields of a propagating EM wave [106].
architectures like homodyne architecture, heterodyne ar- Conventional weather radars used to have single polar-
chitecture, double-sideband architecture, or direct IF sam- ization. They measured three different moments: the radar
pling. These architectures are differentiated based on the reflectivity factor (Z), Doppler velocity (V), and Doppler
frequency/phase of the oscillator and the received signal spectrum width (W). The radar reflectivity factor is the pro-
[96]. Figure 9 shows the block diagram of a homodyme portional power of the signal received, and the Doppler ve-
biomedical radar system. The magnitude of the radar echo locity is determined from the Doppler shift, which is used
signal of a biomedical radar can be given by [100] to measure Doppler width [107]. Single-polarization radars
measure these three moments only with single polarization
nrm v m
S=/A cos U (t - x (rm))(58) as these moments for horizontal polarization will be denot-
m (4r) 2 r m4
ed by Z H, VH, and W H . However, recently, dual-polarization
where r is the range and v is the RCS, n is the attenua- weather radars have been used as they collect data of moments
tion factor, and x is the time taken by the signal to travel for both H and V polarization as Z H, Z V , VH, VV , W H, and W V .
to the human body and back to the radar [101]. Biomedical Clutter involves the unwanted echo that interferes with the
radars are used for human fall detection using an ultrawide- signals reflected by the weather targets. A pulse radar used for
band FMCW radar whose operating frequency is roughly weather monitoring measures the SCR as [108], [109]
5.8 GHz. Based on sudden changes in radar signatures/RCS
S v
data, it is possible to find/predict human fall detection [98], C = v 0 Ri b (cx/2) sec z . (59)
[99]. Radar technology can also be used to make vision
sensors for the visually impaired, which was proved for the And the range of the weather targets is
first time in 2014, and it uses 80-GHz FMCW radar. Today, v
another very important use of radars is the detection of can- R= (60)
(S/C) min v 0 i b (cx/2)sec z
cer, which is a significant threat to human life. The conven-
tional methods of cancer detection like mammography and where v is the target cross section, v 0 is the normalized
magnetic resonance imaging have shortcomings like high clutter coefficient, R is the range of clutter, i b is the azimuth
cost and give harmful radiation. Biomedical radar-based
cancer detection would be a better solution to these prob-
lems [102]. The working principle is the difference between
the dielectric properties of healthy and damaged cancer Transmit
tissues, which can be detected using a radar signal with a Power Antenna
vco 90°
of frequency 3–12 GHz for brain cancer/tumor considering Splitter
0
the safety of the skull and the brain [98]. Phase
Shifter
I(t) VGA LPF
WEATHER RADAR
Weather radars have been of great use during the last cen- LNA
Receive
tury, mainly for measuring precipitation, rainfall, and
Q(t) VGA LPF Antenna
windspeed, and also for detecting the possibilities of hail,
tornadoes, blizzards, and floods. Weather radars have
been in use since the 1950s to detect precipitation, and to- FIGURE 9. The schematic diagram of a biomedical radar. VGA:
day’s weather radar network covers a majority of densely ­variable gain amplifier; LPF: low-pass filter.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 85

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 en-route surveillance. It normally
uses the parabolic rotating antenna
Clock Antenna system. The operating frequency is
Transmitter Duplexer approximately 2,700–2,900 MHz to
Driver Unit
Antenna
achieve the required range, however,
for long-range applications, frequen-
AFC cies must be roughly 1,300 MHz [16],
[111]. The secondary surveillance ra-
dar is a similar rotating radar, but it
is for both approach and en-route
RCU
surveillances, with an operating fre-
quency of 1,030 MHz for ground-
to-air interrogation and 1,090 MHz
Processing
Receiver for air-to-ground reply. Apart from
Unit
ASRs and ARSRs, i.e., ground-based
radars, aircraft have onboard sensors
Beam Angular used for navigation [111]. Airborne
Display Position Sensor weather radars provide data about
the winds and rainfall happening
FIGURE 10. The schematic diagram of a weather radar. AFC: automatic frequency control; around the aircraft. These radars are
RCU: range compensation unit. also capable of detecting turbulence
up to 40 nmi and alert the pilots in
beamwidth, and c, x, and i are the velocity of light, pulse- such situations. A low-range radio altimeter is another ra-
width, and gazing angle, respectively [108]. Figure 10 shows dar-based onboard device used for navigation. It provides
the block diagram of a weather radar. [110]. In addition to data about the terrain beneath the aircraft by using simple
weather parameter monitoring, weather radars are also pulse radars, which serve the pilots in mountainous regions
effectively used for dam-water-level monitoring, dam tilt or unsuitable weather conditions [111]. A scenario of detect-
measurement, river/canal water discharge speed measure- ing/tracking an aircraft by using two independent airport
ment, water-level measurement, gate vibration measure- ground radar systems is shown in Figure 11. These con-
ment, water condor measurement, and so on. ventional methods of detection are not suitable for UAVs/
sUAVs as they are smaller in size and have a very low RCS.
AIR TRAFFIC MANAGEMENT RADAR They also travel at slower speeds, so the Doppler process-
The aviation industry has evolved with a high dependency ing for larger aircraft is not relevant for UAVs [113]. Hence,
on radars. Air traffic has risen with the recent introduc- to increase distance and resolution, UAV traffic sensors use
tion of miniaircraft, air freight cargo flights, and UAVs to higher frequencies and MIMO radars instead of a conven-
this space. Air traffic management radars can be classified tional single-radar system. These UAV traffic systems use
into two types: airport surveillance radars (ASRs) and air FMCW radars with frequencies in the X, K, or W bands de-
route surveillance radars (ARSRs) [111]. Air traffic control pending on the respective airport traffic density [113].
systems use bistatic antennas, which are parabolic and
mechanically rotate continuously for 360°. They transmit SAR
EM pulses at a duty factor of roughly 0.001 and receive sig- SAR is a side-/down-looking radar that uses the principle
nals when not transmitting. Hence, these radars are able to of aperture synthesis to form imagery. It has various ap-
detect the distance, velocity, and direction of the aircraft plications, like foliage-hidden target imaging, military ap-
[16]. There are two levels of air traffic surveillance radars: plications, deformation monitoring, geographical imaging,
primary and secondary surveillance radar. Primary sur- and so on [114]. SAR radars are usually mounted on aircraft
veillance radar covers a large area and is usually used for or UAVs [115] and operated for intended applications. Air-
borne/spaceborne radars are usually tilted toward one side
because if they point perfectly vertically downwards, im-
ages of multiple points on the ground that are equidistant
from the radar will interfere. To avoid this, side-looking ra-
dars are used [116]. Even for a side-looking antenna, there
will be some points on the ground that will be equidistant
from the antenna. If we move the antenna, the Doppler
effect is introduced, and all these points receive different
FIGURE 11. A scenario of tracking an aircraft by two independent signals of Doppler frequencies. Hence, the reflected signals
ground radar systems. will have different wavelengths, thus, with the wavelength

86 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 and time interval data, all the points will be differentiable, ISAR
allowing us to plot a complete image of the targets. Here, ISAR is a technology similar to SAR and is also used for
we are trying to synthesize a longer antenna with a short high-resolution imaging. It is usually used for targets like
antenna by moving the antenna. The range and azimuth cars, ships, airplanes, or spacecraft, which are the targets
resolution of a side-looking radar in the direction of a radar under motion. Aperture synthesis demands relative mo-
beam is given by [118], [121] tion between the antenna and the target. When the an-
tenna is stationary and the target is moving, we use ISAR
cx p
t R = 2 (61) technology [122], [123]. ISAR requires the data of the
targets from their various aspect angles so as to combine
m
t Az = L R (62) them to construct the images on a range and cross-range
Az
plane. The main difference between the SAR and ISAR sys-
where t R is the pulse length, c is the speed of light, x p is tems is that the radar system moves in the SAR imaging
the time gap between the transmission of two consecu- while the targets move in the ISAR imaging. The collected
tive pulses, t Az is the resolution in the azimuth direction, data are processed to finally obtain a 2D ISAR image. The
m is the wavelength, L Az is the length of the antenna, and range Doppler algorithm is used to process ISAR images
R is the slant length or the slant distance between the an- because it provides better range resolution [15]. As it is ca-
tenna and the ground [117]. Figure 12 shows the block pable of functioning during the day and at night and in
diagram of an SAR system, with the motion measurement all weather conditions, ISAR has a couple of uses to de-
sensor being the most critical part. Three popular modes tect moving targets. It is used mainly for maritime sur-
of SAR are shown in Figure 13; therein, the stripmap SAR veillance to detect ships and vessels in the ocean and for
[refer to Figure 13(a)] has a side-looking antenna, the ra- imaging space objects that are moving.
dar beam footprint moves as the radar moves in the form
of a strip (the width of this strip is called the swath), and GPR
the images are formed based on the strip-on-the-ground GPR is another application of radar that is used mainly
data collected [118], [119]. The spotlight SAR [refer to Fig- for the detection of subsurface buried or hidden targets
ure 13(b)] is adjusted such that it spotlights the same area by penetrating EM waves through the ground [124]. It
as it moves, and the Doppler center
of the reflected signal varies linear-
ly with time [119], and the sliding
Pulse
spotlight modes or terrain obser- Transmitter
Generator
vation by progressive scans [refer
to Figure 13(c)] are similar to the Motion
spotlight mode, but it does not spot Measurement
Radar Control Duplexer
Sensor Antenna
light at a single point. This method
improves the resolution of the im-
ages. The Doppler center of the re- SAR Image
ADC Receiver
flected signal varies linearly with Processor
time, and the azimuth spectrum is
aliased [119], [120]. FIGURE 12. A top-level schematic diagram of an SAR system.

(a) (b) (c)

FIGURE 13. Different operating modes of SAR systems. (a) A stripmap SAR, (b) spotlight SAR, and (c) sliding spotlight or terrain observation
by progressive scans.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 87

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 has several uses in many fields, like geophysical applica- are transmitted on the target and the backscattered waves
tions (studying soil characteristics), water-level detection, are received. Using the data brought by these waves, the
mining activities, metal detection, fiber-cable detection, target can be detected/processed. However, GPR has some
pipeline detection, analysis of ice caps, antipersonal mine challenges and problems, like identification of the back-
detection, and so forth. The working principle of GPR is scattered signal from the data collected by the receiver
simple, like a normal radar [124], [125] where EM waves [125]. Figure 14 shows the basic block diagram of a GPR
system [126]. GPR uses several configurations of anten-
nas, the most popular of which is the multibistatic con-
figuration, wherein the bistatic antennas move together,
Source and Display/
Modulation Storage maintaining the relative position of both the transmit-
ting and receiving antennas. Other configurations are the
multimonostatic configuration, where a single antenna
Power Signal is used, or the two antennas are placed very closely to
Amplifier Processing each other with respect to the wavelength transmitted,
and the common midpoint configuration, where both an-
tennas steadily move away from each other, maintaining
LNA the required spacing [125]. GPRs usually use a frequency
of 0.5–2 GHz. The different types of GPRs are forward-
looking GPR, side-looking GPR, and airborne/drone-
Transmitter Receiver mounted GPR [124].
Antenna Antenna
OIL SPILL DETECTION RADAR
Target Recently, oil spills in oceans have become a serious issue
as they cause extensive damage to the marine ecosystem
and waste resources that are important to human civili-
zation. To prevent these oil spills and avoid such dam-
FIGURE 14. The schematic diagram of a GPR system. age, proper identification and inspection of these oil spill
sites is necessary [115]. There are several methods used to
identify these sites, two of the most popular being aerial
Extract photography in the visible or near-IR spectrum, and SAR
SAR Isolate Dark
Statistical
Image Areas imaging. Oil has different optical properties for reflec-
Characteristics
tion and absorption, enabling us to differentiate it from
water [128]. However, some other phenomena present in
Statistical
Comparison
Oil or Lookalike oceans backscatter SAR signals in a similar way to those
Database Decision backscattered by oil spills, and these phenomena make
distinguishing water from oil a critical task. Usually, SAR
FIGURE 15. A methodology for oil spillage detection. is used to detect oil spills in the C band, but it does not
give an accurate thickness of the oil spill [129]. Figure 15
shows a methodology used to detect oil spills. In an oil
spill detection process, the first step is preprocessing the
image, which involves land masking, and then identify-
ing the dark spots in the SAR image. Factors like weather,
ship routes, and wind speed influence image quality, so
all these factors and their statistical data are considered to
differentiate oil spills from water [130].

FOPEN RADAR
FOPEN radar is the most important method used for de-
tecting and imaging the targets hidden under tree can-
opy and foliage, as shown in Figure 16. Generally, SAR
systems are used for FOPEN applications [131]. FOPEN
SAR systems usually work in ultra HF, VHF, L bands, or S
bands, despite poorer resolution at low-frequency bands,
but give long-distance penetration in foliage environ-
ments. This is because higher-frequency waves are inca-
FIGURE 16. A working illustration of an FOPEN radar. pable of penetrating through foliage and canopy [132].

88 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 FOPEN radars often give false alarms due to the presence this would each be one half of the elements, divided in the
of tree trunks, leaves, bushes, and so forth. However, a middle. For a planar array, these subarrays would be the
solution to this can be the effective use of radar-signals four quadrants of the array, each with one quarter of the ar-
polarization. As tree trunks are vertical, vertically polar- ray’s elements. The angle difference Di in the phase-based
ized waves exhibit more attenuation than horizontally monopulse tracking/guidance is calculated as
polarized ones [133].
2d sin (i)
Di = (65)
TRACKING AND GUIDANCE RADARS m
Tracking and guidance is an emerging remote sensing ap-
plication that uses radar technology to improve efficien- where m is the radar wavelength, d is the distance between
cy in military activities by accurately tracking known/ two adjacent antennas, and i is the aspect angle of the ra-
unknown targets and effectively guiding friendly targets. dar. The Kalman filter-based tracking radar, another popular
Also, the capability of precisely updating the path/trajec- standard tracking and guidance radar/technique, uses Kal-
tory of missiles, the tracking and guidance system sup- man filters that calculate the
port for improving the accuracy of attack, and reducing position, acceleration, and ve-
the overall mission’s accomplishment cost are benefits locity of a target. Based on past
of these radars. There are several standard tracking and measurements, the recursive THE RANGE DOPPLER
guidance radars/techniques available, with the most Kalman filter algorithm pre- ALGORITHM IS USED TO
popular being monopulse tracking radar, which uses dicts the target’s current state. PROCESS ISAR IMAGES
four antenna beams to follow the angle of arrival (AoA) Kalman filter-based track- BECAUSE IT PROVIDES
of a target. To do this, a comparison is made between ing radars are more common BETTER RANGE
the amplitude of the radar-received signals from each because they can account for RESOLUTION.
antenna beam/quadrant. Monopulse tracking radar uses noise in radar measurements.
the amplitude/phase differences to calculate target dis- They also resist environmen-
placement from the boresight of the antenna plane and tal effects and interference
the error information calculated, as in the following, to from other radar systems considerably. The state-prediction
steer the tracker: equations in the Kalman filter-based tracking and guidance
radar system are
ie = c m (63)
Di 1 Di 2
*
Di m D i m
x (k) = F (k - 1) x (k - 1) + B (k - 1) u (k - 1)(66)
where i e is the tracking error; i 1 and i 2 represent the phase P (k) = F (k - 1) P (k - 1) F (k - 1) T + Q (k - 1)(67)
difference between the sum and sum-difference channels,
respectively, and i m is the smallest phase shift that can be where x(k) is the predicted state vector at time-step k, F(k–1)
detected. Another popular standard tracking and guidance is the state-transition matrix, B(k–1) is the control input
radar/technique is conical tracking radar, which uses a ro- matrix, u(k–1) is the control input vector, P(k) is the pre-
tating antenna beam to measure the target’s AoA. The an- dicted covariance matrix, and Q(k–1) is the process noise
tenna beam rotates at a constant speed, and the target’s AoA covariance matrix. Another popular standard tracking and
is defined by the phase of the radar signal received from the guidance radar/technique, AI-based radar target tracking,
target. The tracking error in the conical tracking technique uses ML intelligent algorithms to track/monitor/guide the
is calculated as targets in radar applications. AI-based radar target-tracking
systems frequently employ a combination of signal process-
ie = c * sin (i t) m (64)
Di p ing, ML, and AI technologies to track/guide targets. Such
Di m algorithms might be used to learn the target’s mobility
patterns and predict its future position. This information
where i e is the tracking error; i p represents the phase dif- is then utilized to properly track the target. Reinforcement
ference between the sum and sum-difference channels, learning, you only look once, support vector machines,
respectively; i m is the smallest phase shift that can be de- deep neural networks, DCNNs, and several more AI al-
tected; and i t is the target’s angle. Yet another popular stan- gorithms are recent ML/DL techniques used for AI-based
dard tracking and guidance radar/technique is phase-based radar target tracking. A cognitive radar-tracking algorithm
tracking, wherein the main remote sensing technique is is another popular standard tracking and guidance radar/
a monopulse phase comparison that typically measures technique. It can adjust itself to changes in accordance to
the phase difference/error information from the displaced the surroundings as well as the target’s behavior. Tradi-
phase centers of an array antenna, unlike the amplitude tional tracking algorithms, on the other hand, have been
comparison. In the phase comparison, the monopulse ar- created to follow targets in a known and unchanging envi-
ray is subdivided into subarrays, and then a total sum and ronment. Cognitive algorithms can follow targets in a range
a difference or del channel are formed. For a linear array, of adverse environments, such as those with many targets,

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 89

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 in fog, and through interference. The adaptive waveform tasks, such as weather monitoring and missile tracking. A
selection method is one example of a cognitive algorithm tracker is often used to monitor radar targets and uses radar
for target tracking and guidance. This technique uses ML measurements to determine the target’s condition, such as
to decide the best radar waveform for tracking a particular location, velocity, and acceleration. The tracker then uses
object. The algorithm considers the target’s motion, and it these data to calculate the target’s future location. This pro-
cognitively tracks the target as cedure is done continuously to follow the target in real time.
Radar target scheduling is the process of distributing radar
PT = f (R, C, I, W)(68) resources to various activities, such as tracking, searching,
and weather monitoring. The purpose of radar scheduling
where PT is the probability of tracking the target, R is the is to improve a radar system’s overall performance. Time-
range of the target, C is the cluttered environment, I is the constrained scheduling can be given as
interference environment, and W is the cognitive algorithm.
Real-time radar target tracking, another popular standard maximize sum (x i)(69)
tracking and guidance radar/technique, is useful for track- subject to sum (x i) 1= T (70)
ing targets in real time. FPGAs, for example, can do the nec-
essary computations quickly, where x i is the dwell time for target I, and T is the total
and several effective tracking available time. The two main techniques that are essential
algorithms can run in parallel to achieve accurate tracking and guidance for a missile’s
RADAR TARGET in it. There are multiple meth- path are the guidance algorithm and the missile’s autopilot,
SCHEDULING IS THE ods for real-time FPGA-based in which the commands are sent from the guidance algo-
radar target tracking. These rithm to the missile’s autopilot, which guides the missile to
PROCESS OF DISTRIBUTING
algorithms are usually well reach the target [134]. The working of the tracking and guid-
RADAR RESOURCES TO
optimized and implemented ance system for a missile guidance mission can be divided
VARIOUS ACTIVITIES, SUCH
in the hardware’s high-speed into several phases. In the first phase, the target is fixed
AS TRACKING, SEARCHING, platforms, such as FPGAs, and this information is transferred to the subsystems of the
AND WEATHER Jetson Nano Nvidia, GPUs, missile. The next phase is a quick one in which the missile
MONITORING. and so on. A more general- is launched and it starts the homing flight. The guidance
purpose tracking algorithm system consists of a seeker, which provides data about the
implemented on an FPGA is angle and range of the target, and the inertial instruments,
a different approach for re- which provide the body rates and parameters of the motion
al-time multiple self-adaptive radar target tracking. This of the missile. Figure 17 shows a block diagram of a missile
technique is more complex, but it is also easier to adapt and guidance control system. In the same way, the autopilot sys-
allows for the use of more advanced tracking algorithms tem carries the missile closer to the target. Once the missile
running in one hardware platform, and the appropriate one is within a specific range from the target, the seeker takes
will be enabled whenever it is required. Radar target track- over and guides the missile to the homing phase. The seeker
ing and scheduling, another popular standard tracking and provides information about the position of the target, and
guidance radar/technique, is the process of continuously this information, along with other available data, is passed
monitoring targets using radar and giving radar resources to the filter. Then, the autopilot system decides the trajec-
to maximize tracking performance. This is a difficult task tory to be followed. The final phase is the end game, during
as radars must balance the requirement to track many tar- which the missile and the target are in close proximity. In
gets with the need to find new targets and conduct various this phase, the target attempts to escape with high accelera-
tion, and the seekers starts capturing incorrect information.
The seeker and guidance system are then switched off to
Target avoid interruptions with last-second instructions. The mis-
Dynamics
sile either kills the target or misses it [135].
There are several methods available to provide hom-
Missile Counter- ing guidance to missiles, with the proportional navigation
Dynamics measure
technique being the most popular one. Proportional navi-
gation is considered a standard guidance principle that
Autopilot Sensors Seeker works on the basis of angular rate change and LoS. It states
that the rotation of missile velocity vector and the rotation
of LoS should be proportional [136]. Some other meth-
Autopilot Guidance
System ods of homing are active, semiactive, and passive homing.
Active homing uses active radars that can independently
FIGURE 17. The schematic diagram of a missile guidance sense the target and guide the missile toward it. Semiac-
control system. tive homing uses passive radar and a separate transmitter

90 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 to illuminate the target, while passive homing also uses down conversation units, amplifiers, and power dividers.
passive radar with radiations originating from the target. The basic idea of SDR is to replace these physical components
It can also sense the target with an IR sensor that detects with software elements that will perform the tasks of these
the heat generated by the target [137]. physical components. SDR has several advantages over con-
ventional radars. A radar implemented using SDR can serve
MODERN STARING RADAR SYSTEMS various purposes, including making advanced signal pro-
In this section, we delve into modern staring radar technol- cessing techniques more feasible compared to conventional
ogies, which include MIMO radars, SDRs, cognitive radars, radars. A radar system implemented with SDR becomes re-
photonics radars, quantum radars, noncooperative sensing, usable, runtime reconfigurable, faster, and cheaper [143],
and the Internet of Radars (IoR). [145]. Implementation of an SDR is simple because the plat-
form has all the necessary onboard components/ICs, such
MIMO RADAR as an RF front end, ADCs, DACs, digital signal processors
Radar technology aims to remotely detect a target and its (DSPs), and FPGAs, unlike classical radar systems in which
unknown parameters, like its range, speed, activities, and all these components are required to be connected using the
coordinate. One of the multiple approaches available to do proper cables/data bus. FPGAs are ICs that are configurable,
this task is directing a beam of EM waves toward the target which makes them capable of controlling radars and pro-
and processing the return signals. For a particular direction, cessing outputs. Figure 19 shows the basic block diagram of
having an array of antennas instead of a single antenna, an SDR platform [146]. SDR is widely used and becoming
supports phased-shifted transmission and reception. Beam- more popular due to all the advantages it offers over a con-
forming is the process where an array of closely spaced ventional radar. It has applications in the defense and space
transmitter/receiver antennas can steer an EM beam in any sectors and has also been put to commercial use. The idea of
direction in the space. Beamforming has several advantages, having cognitive radar is possible only because of SDR radar
including that no mechanical components are involved in and its FPGA-based implementations [144].
steering the EM beam, so that the system can be electroni-
cally controlled. Thus, performance improvements are pos- COGNITIVE RADAR
sible just by improving on the advanced signal processing Cognition, as defined by the Oxford Dictionary, is the men-
techniques used in terms of phase variations of all the re- tal action or process of acquiring knowledge or understanding
ceived signals from the ­array of the antennas [138]. through thought. It is an activity humans perform most of
As mentioned previously, the steering of RF beams the time. We do it subconsciously, and that is what dif-
is electronically controlled, which makes it possible for ferentiates us from other species. It involves activities like
the transmitting array to steer the beam [139]. Electronic- thinking, reasoning, learning, and decision making [148].
steered arrays (ESAs) steer by introducing or changing the But that is about humans. How can an engineering system
amount of phase shift in the transmitted waves [140]. A be cognitive? With the recent advancements in ML, the
block diagram describing the functioning of this is shown integration of engineering systems and ML computer al-
in Figure 18, which led to the development of the MIMO gorithms make engineering systems comparable to cogni-
radar system, and the same principle is also used in MIMO tive/smart systems, but the clear definition of cognitive/
communication systems. It significantly reduces variations smart systems is still hazy [150]. Adaptive radars change
in the SNR. The target is detected by
observing the change in the RCS.
More than 10-dB scintillations are Arbitary Transmit
observed for small changes in the an- Waveform Antennas
Generator
gle [138]. MIMO radars demand dif- Filter Amplifier
ferent coding and signal processing Power
PSG Coupler Splitter
techniques compared to convention-
Filter Amplifier
al antennas. The data collected by
multiple antennas are combined to
achieve better output. With the help IF Filter LNA Filter
of multipath channels, MIMO radar
is capable of exploiting spatial prop- PC ADC IF Filter LNA Filter
erties, which makes it more capable
than conventional radars. It also has
IF Filter LNA Filter
more degrees of freedom [141], [142].
Receive
Antennas
SDR
SDR involves various hardware parts, FIGURE 18. A schematic diagram of an MIMO radar. IF: intermediate frequency; PC: personal
like signal generators, filters, up/ computer; PSG: programmable signal generator.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 91

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 the methods of processing data as a function of time. But PHOTONICS RADAR
there is no change in the actions performed by the radar. Conventional radar, which usually operates in mm-wave
Information flow during the function of adaptive radars bands, requires a wide bandwidth, for which today’s
is one way how radars send the initial signal. Cognitive fiber-optic networks have been a solution to this as they
radars aim for adaptive hardware, analytical techniques, make transmission in the optical spectrum easier and
and involving cognitive processes in their function more efficient. However, using optical networks requires
based on technologies like ML, AI, and stochastic con- high-modulation frequencies, detection, and speed of the
trol [150], [151]. The idea of transmitted signal. Optical networks require optical sourc-
cognitive radar was first in- es, high-speed photo detectors, and optical signal modula-
troduced by Simon Haykin tors [154]. Photonics radar aims to replace all the electronic
QUANTUM RADAR IS ONE [153]. Cognitive radar can systems in a conventional radar by a photonic system. It
THAT IS SIMILAR TO A be realized by using a de- involves subsystems that convert electric signals to optical
cision-theoretic approach. signals and vice versa [155], [156]. This radar is different
CONVENTIONAL RADAR,
The radar collects data from from lidar and conventional radars. To transform electri-
BUT IT WORKS ON THE
the environment based on cal signals to optical signals, special types of lasers, like the
PRINCIPLES OF QUANTUM
which decisions are made mode-locked laser and distributed feedback laser, are used.
MECHANICS, NOT with respect to the decision- Figure 20 shows the schematic diagram of a photonic radar
CLASSICAL MECHANICS. theoretic approach. A prin- system. The optical signal is converted to an electric sig-
cipal component for this is nal with the help of photodiodes [158]. The optical beams
passive environmental sens- produced are modulated by a direct laser modulation or
ing, which collect all other an external modulation. Modulators like Mach–Zehnder
data than target detection [152]. It will form a required modulators (MZMs) are used for the same purpose [157].
knowledge database of cognitive radar, and this will be In a conventional radar, the signals produced and received
put to use by learning from the available database. For are transferred through coaxial cables or waveguides,
cognitive radar, the receiver to the transmitter is another which are large, bulky, and costly. They are also poor at re-
principal component [149], [148]. sisting interference and have high dispersion and low ca-
pacity. Optical signals, on the other
hand, can be sent simultaneously in
Signal Signal Translation RF both directions with different wave-
Processing and Conversion Front End lengths for different directions. This
reduces the size, weight, and cost,
ADC LNA and, at the same time, resists interfer-
ence and provides less attenuation.
Such a radar that uses laser and EM
Quad components is called a photonic-based
Duplexer
DSP DDS fully digital radar system [154], [159].
Antenna

QUANTUM RADAR
Quantum radar is still an emerging
DAC PA technology and attempts are being
made to idealize and put it to work,
but there is still a lot of scope for fur-
FIGURE 19. The schematic diagram of an SDR platform. PA: power amplifier. ther research in this field. Quantum
radar is one that is similar to a con-
ventional radar, but it works on the
Radar Transceiver principles of quantum mechanics,
not classical mechanics. The main
Photonics- Digital
Mode-Locked Photonics- phenomenon used in quantum radar
Based RF Signal
Laser Based ADC is quantum entanglement [160]. If
Generator Processor
there are two noninteracting objects,
then, according to classical mechan-
RF Path
Optical Path ics, the measurement of states of one
Digital Path of the objects is independent of the
measurement of states of the other
object. However, in quantum me-
FIGURE 20. The schematic diagram of a photonics radar system. chanics, the measurement of states of

92 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 one of the objects could depend on
the other. This phenomenon is called Entanglement
Generation Transmitter
quantum entanglement. Seth Lloyd
came up with the idea of using this Signal T
phenomenon to make a new kind of Photon A
R
radar called quantum illumination ra-
G
dar (QIR), which, today, is often re- E
ferred to as quantum radar. The other T
approach used to make quantum Receiver
radar is quantum interferometric
radar, which gives results by measur-
Idler
ing phase differences [161]. Figure 21 Photon
shows a principle illustration of Entangled
quantum radar [162]. The working of Photon-State
the QIR is based on the entanglement Processing, Estimation, Target Information
Quantum and Information
phenomenon, as mentioned previ- Memory Extraction
ously. It starts with the creation of a
pair of correlated entangled photons. FIGURE 21. The principle illustration of a quantum radar.
One of them is the idler photon, and
other is the signal. The quantum state
of the idler photon is stored in quantum hardware while the of noncooperative target detection radars include defense
signal photon is transmitted in space. If this photon meets surveillance systems, evaluating the flight state of a faulty
the target, it is reflected back to the receiver of the quantum satellite, generating target data for missions related to space
radar. The receiver continuously receives photons, which debris, safe drone landings, unmanned traffic management
are either noise photons or reflected-signal photons. To dif- systems, detection of stealth aircraft, determining the point
ferentiate these received photons, they are compared to the on the ground targeted by a remote sensing satellite’s lens
idler photon. Signal photons are eventually identified with through attitude estimation, and so forth [165].
the help of entanglement correlations.
IOR
NONCOOPERATIVE TARGET-SENSING RADAR In a general context, the Internet of Things is about a set of
A noncooperative target is one whose state or motion is not devices that are interconnected and constantly exchanging
known. A malfunctioning satellite or space debris in space
applications, missiles or enemy aircraft in defense applica-
tions, spy birds/UAVs for airborne applications, and colli-
sion detection and avoidance of targets/objects could act as Navigation State
noncooperative targets [163]. High-resolution range profile
(HRRP)-based remote sensing or an identification system
has to be used for detection of noncooperative targets. Usu- Searching Obstacle
ally, an active electronically scanned array radar system is Communicate to
used for this purpose [164]. An HRRP is a 1D radar image Computer
that projects the reflectivity of the target into the LoS. The Detecting Obstacle
type and orientation of the target is identified based on
different HRRPs, however, they have several technical is-
sues, and some of them can be avoided by working in the Comparison Maneuvering for
frequency domain. The phase of HRRPs decides the loca- to Criteria Avoidance
tion where the HRRPs of the target appears in the range
window, which means that the range profile of the target
can be anywhere in the range window. The appropriate Tracking Obstacle
corrective measures need to be implemented to avoid such
Reaction Bearing
potential problems. One interesting application of such
radars is collision-avoidance radar, in which noncoopera- Range,
Velocity,
tive target detection radars are mounted on a UAV. These Bearing Reaction Time
radars identify flying obstacles and change the path of the
UAVs to avoid collision. An algorithm for possible collision
detection and avoidance in a noncooperative target sce-
nario is shown in Figure 22 [165]. The main applications FIGURE 22. A noncooperative target sensing algorithm.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 93

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 the data they acquire. With all the technological advance- methods to solve radar and communication interfer-
ments, radars are finding various applications in our day- ence mitigation and spectrum management solutions
to-day lives, like radars on cars, radars on drones, or even using different ways like oppor tunistic spectr um ac-
radars on phones, and radars on wearables. Interconnection cess, interference channel estimation, and optimal
and exchange of information among all such radars and receiver utilization/design. Co-design aims to perform
existing global communication networks become signifi- the best operation for dual-function radar communica-
cant, which is being attempted in the IoR. The IoR can have tion by finding unique radar remote sensing waveform
various applications, like personal and home applications, sets. Rad-com involves techniques like information-em-
applications in health-care radar, vehicles/drones, and so bedded waveform diversity and symbol construction via
forth. This can help to have a more cognitive system with transmit beamforming. Com-rad, similar to an OFDM-
multiple radars. Mainly, the IoR uses a technology called based system, merges radar sensing and communication
joint radar communication (JRC), which has several advan- operations. An architecture showing the IoR is presented
tages like less energy consumption, lower cost, and reduced in Figure 23 and connects an array of radars to the main
spectrum usage. It enables concurrent operation by estab- server via the available Internet network, i.e., the inter-
lishing communication links while sensing the targets via continental exchange-application programming interface
radar processing. (ICE-API). The radar measurements are routed through
The existing types of JRC are coexistence, co-design, multiple stages of residue-concatenate fully convolutional
communication/control utilized by radar systems (rad- network processes/standards. Emerging IoR research also
com), and radar sensing utilized by a communication supports reducing (off-loading the computational cost
scheme (com-rad). Coexistence deals with finding new at the radar station) computational burdens at the radar
stations, meaning that the radars’ remote sensing data
are transferred via the ICE-API network into the remote
Radars server of the high-end configuration in which all the main
signal processing and AI-based classification engines are
running. This kind of IoR-supported radar remote sensing
permit data fusion, the simultaneous capturing of targets
from different angles, and simultaneous handling of mul-
tiple targets that enable the accurate construction of SAR
and ISAr images.
Authentication
EW SYSTEMS
EW is the use of the EM spectrum for defense applications
ICE Client to extract information about the enemy’s surveillance sys-
tems, attack the enemy’s electronic system, and manipulate
RC-FCN_Train the enemy’s use of the spectrum while protecting one’s own
Receive: Train information use of the spectrum [166], [167]. The basic idea of EW is
RC-FCN_Test to restrict access to the EM spectrum while it is being used
#Train has stopped for military purposes [166]. EW techniques are commonly
RC-FCN_ GetTrainLog divided into three categories: electronic support measures
Receive: Train log
(ESMs), electronic attacks (EAs), and electronic protection
(EP) [see Figure 24] [166], [167], and the briefs on these
three EW techniques are given as follows.

ESMS
ICE–API ESMs or EW support is the gathering of EW data through
electronic intelligence (ELINT), communication intel-
ligence (COMINT), and the ESM receiver. They involve
passive radars used for identification, analyzing, and lo-
cating enemy radars for threat detection. Figure 25 shows
the general schematic diagram of an EW system. An ESM
High-Performance is responsible for collecting the information required
Server Database
Computer for electronic countermeasures (ECMs) and electronic
counter countermeasure (ECCM) operations [166]. The
ELINT system is used to obtain and compile data about
FIGURE 23. The IoR architecture via the intercontinental exchange- the weapons and electronic systems operated by the en-
application programming interface (ICE–API) network. emy. The COMINT system is the intelligent system based

94 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 on the communication happening
with persons other than the intend-
EW System
ed recipients. Together, the ELINT
and COMINT systems are called
the signal intelligence (SIGINT) sys-
tem. Both ESMs and SIGINT are re- ESM ECM ECCM
sponsible for the passive collection
of data, but ESMs are used for tac-
tical purposes with immediate ac- Active Passive
Spatial
tion, while SIGINT is used to collect ECM ECM
intelligence data, with a subsequent Spectral
analysis [166].
Noise Chemical
Jamming ECM Temporal
EA
An EA, also known as an ECM, in- Deceptive Mechanical Netting
volves the actions taken to disturb Jamming ECM
the normal operation of the enemy’s
radar. ECMs are classified as active FIGURE 24. A top-level classification of EW systems. ECM: electronic countermeasure; ECCM:
and passive. The transmission of EM electronic counter countermeasure.
waves to make the functioning of the
enemy’s EW less effective is an active
ECM. It involves noise jamming and deceptive jamming and netting [166]. Spatial ECCMs involve methods based
[167]. Noise jamming prevents the radar from identifying on space, like side-lobe cancelers, side-lobe blanking [166],
the target by creating disturbances, which masks the signal [167], and the burn-through technique. Spectral ECCMs, as
of interest. Different types of noise jamming are spot jam- the name suggests, are based on frequency. Methods like
ming, sweep jamming, and barrage jamming. In spot jam- low probability of intercept, frequency agility, and Dop-
ming, the jammer is concentrated at a particular narrow pler filtering are used for spectral ECCMs. Temporal ECCM
band of frequencies. In sweep jamming, the jammer has a methods, like pulse-expansion contraction [166], [167], PRF
wide bandwidth and sweeps through all the frequencies. In agility, and constant false-alarm rate, are time dependent.
barrage jamming, multiple frequencies are simultaneous-
ly used to overcome the problems of spot jamming [166], PHOTONIC EW SYSTEMS AND SIGNAL PROCESSING
[167]. Deceptive jamming involves deliberately creating This section discusses an emerging remote sensing tech-
false signals to mislead the hostile electronic system. Differ- nology that uses photonics technology in microwave ap-
ent ways of deceptive jamming are false-target generation, plications, i.e., photonics for EW systems and microwave
range deception, velocity deception, and angle deception. signal processing.
False-target generation is used to generate false signals for
the victim’s radar, which confuses the enemy’s radar. In
range deception, the jammer locks the target and generates
pulses to confuse the radar about the actual position of the
target. Velocity deception is used for CW and Doppler ra- Jammer,
Detect Signal
dars, which track the targets based on velocity or Doppler Activity Chaff,
frequency shift [166]. Passive ECMs do not have their own and so on
sources, instead they use confusion reflectors to disturb
function of the enemy’s radar. This is done using chemi- Deinterleave Measure
Angle, Freq, ECM Action
cal or mechanical means. Chemical agents, like smoke and Signals
PRF
aerosol, are used for chemical jamming, while mechanical
jamming involves the use of mechanical objects like drones
Determine Signal Measure
or flares to disturb function of the enemy’s radar [166]. Type and Angle, Freq, Action
Characteristic PRF
EP
EP is also known as ECCMs. The main function of ECCMs Determine ELINT Display,
is to ensure effective working of the EW system by overcom- Results
Location Database Database
ing the effects of the enemy’s ECM [168]. These techniques
are based on fundamental properties of the radars, like
power, frequency, pulse length, PRF, gain, polarization, and FIGURE 25. The schematic of a general EW system’s architecture.
so on. The types of ECCMs are spatial, spectral, temporal, Freq: frequency.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 95

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 MICROWAVE PHOTONIC the remote antenna’s location used signal processing
EW SYSTEMS techniques like video detection or down conversation to
The main idea of microwave photonic system is to modu- tackle such problems earlier. However, MPLs have prov-
late the radio signals with optical carriers and transmit en to be a more effective solution by reducing the weight,
them through fiber-optic links for photonic signal pro- size, transportation loss, and power consumption at the
cessing. One intuitive example of microwave photonics remote location [170].
is the microwave photonic link (MPL). As mentioned
before, broadband RF energy undergoing amplitude PHOTONIC SIGNAL PROCESSING
modulation on an optical carrier source is the working A conventional signal processor has limitations on the
principle of MPLs [169]. In such MPLs, photo detectors sampling speed it can perform. A photonic signal proces-
are used as terminal points where a modulated electric ser solves this problem by working at high-multigigahertz
signal is obtained back from the incident optical signal. sampling frequencies, which provide a better processing
Figure 26 shows the block diagram of a simple MPL. The speed [171]. Broadband RF signals are modulated us-
EW system demands the transmission of analog signals ing optical carriers and processed using optical devices.
for which the modulation of photonic links can be done The architecture for a simple photonic signal processing
either internally or externally. The laser current is driven is shown in Figure 27. A photonic signal processor aims
by a modulating RF signal, which is known as internal to work the same as an ordinary microwave filter in an
modulation, while external modulation is done using an RF system. It does the processing of higher bandwidth
interferometric device [170]. Transmitting a broadband signals by taking advantage of optical delay lines [171].
signal using the fiber cables over long distances has sev- A photonics signal processor has several advantages over
eral advantages. Direction finding requires a long base- a conventional signal processor, like lower losses, higher
line technique, which uses many EW receiver systems, bandwidth, and immunity to EM interference. It is also re-
thus, all the RF signals have to be transported to a com- configurable and fast tunable, which cannot be achieved
mon processor module. The phase shift is measured be- with traditional electrical signal processors [172]. Photon-
tween the signals received by separate antennas located ic signal processing starts with an RF input signal, and the
far apart [169]. A longer antenna separation is possible signal has sidebands with the center frequency ! fRF . This
due to the MPLs, which gives freedom to the EW sys- input signal may come from an RF generator or the anten-
tem designer to position the antennas. Traditionally, nas. The signal is passed to the modulator and gets con-
verted to the optical domain, with
the center frequency of the spec-
trum sidebands being v ! fRF, where
Laser Modulator the optical source driving the mod-
Modulator Photodiode
Diode Bias Controller Long Fiber-Optic ulator has frequency v. An optical
Cable signal processor, consisting of sev-
Dc Control eral photonic devices, processes the
RF Input Loop RF Output
signal. A photodiode or any optical
receiver is used to convert the signal
FIGURE 26. A schematic diagram for the transportation of an RF signal over an optical carrier.
back to the electrical domain. This
processed signal can be received by
an RF receiver, or it can be reradiat-
ed by passing it to the antenna based
on the applications [172].

PHOTONIC ADC
A conventional electric ADC pri-
RF –f RF 0 f RF –f RF 0 f RF RF
Generator Receiver
marily performs three tasks: it sam-
ples the analog signal at a particular
frequency, holds the corresponding
Laser Optical System voltage for a minimum time, and
Modulator Photodiode
Diode (Photonic Devices) then quantizes the voltage into lev-
els [173]. The inverse of sampling
frequency is usually more than the
ν–f RF ν ν+f RF
time required for voltage sampling;
ν–f RF ν ν+f RF
a photonic ADC also intends to
do the same. One might think of
FIGURE 27. The schematic diagram of a simple photonic signal processor. a photonic ADC as a device that

96 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 converts optical analog signals to optical digital signals, but it uses an optical technology/carrier [178]. This makes
but a photonic ADC, just like a conventional ADC, con- it faster and more accurate, improving its efficiency at re-
verts electrical analog signals to electrical digital signals mote sensing and data collection. Lidar uses coherent light
[174]. Figure 28 shows the block diagram for a photonic energy within a very narrow beam, providing pulses of
ADC [175]. In general, a photonic ADC can be divided high peak intensity, and the intensity varies with respect
into four classes: photonics assisted, photonics sampled, to the sensing/imaging distance. Commercial lasers used
photonics quantized, and photonics sampled and quan- to make the lidar have frequencies of roughly 200 KHz
tized. In photonics-assisted radar, both sampling and and used inertial measurement units to improve pointing
quantization happens in the electric domain, and pho- precision to millimeter levels[178]. The working principle
tonics-based components are used just to optimize the of a lidar system is shown in Figure 30(a). The optical
performance. In photonics-sampled radar, the sampling source transmits the light via a transmission telescope in
is done in the photonics domain, while quantization a well-defined profile (peak power, average power, wave-
is done in the electrical domain. Photonics-quantized length, polarization, divergence, direction, and so forth),
radar does the opposite by sampling in the electronic falls on the targets, and bounces back to the receiver tele-
domain and quantizing in the photonics domain. Pho- scope, which processes it and constructs an image of the
tonics-sampled and quantized radar does both processes target. The structure of the target exposed to the optical il-
in the photonics domain [173]. lumination produces the time delay, phase difference, and
frequency shift, which are optically recorded and used to
RADIO OVER THE FIBER formulate the image in the lidar [22]. Lidar can be oper-
The idea for radio over the fiber (RoF) came in the early ated in different ways, attaching it with an airborne plat-
1980s in RF subcarrier multiplexing, which was used form, a terrestrial tower, vehicle, or fixed ground point for
for distributing cable TV channels. Figure 29 shows the different applications. Today’s lidars are actively engaged
topology of the RoF scheme. The early attempts at RoF for mapping landforms, ocean beds and disaster man-
aimed to provide a distribution method for mm-wave agement, autonomous cars, infrastructure mapping, and
signals over optical fiber to solve the problem of spatial many more applications [179].
congestions in the lower microwave region. In this way,
RoF technology takes advantage of both wireless (i.e., ra- OPTICAL INTERFEROMETRY-BASED
dio over free-space optics) and fiber-optic methods [176]. TARGET SENSING
Initially, the creation and transfer of mm-wave signals Optical interferometry is a technique that studies the
in the optical domain remained the target of research in interference of two or more optical images/scenes of tar-
this field, with the belief that it was the future of broad- gets and constructs the images based on the differences
band wireless technology. But today, RoF is the technol- between those images [17]. Interferometry is used by sev-
ogy where a radio signal modulates the light transferred eral radar sensors, including an optical SAR system. Opti-
through an optical link [177]. Conventional optical cal interferometry can provide 3D data about the image
systems use digital data transmission, while RoF uses by comparing 2D images [95]. A working principle of an
analog transmission. This offers several advantages, like
huge bandwidths up to terahertz.

Analog RF
OPTICAL TECHNOLOGY FOR REMOTE SENSING Signal
Apart from radar-based remote sensing, optical technolo-
gy also plays a vital role in remote sensing. In fact, optical-
based remote sensing technology would give much better/ Mode- Electro-Optic Power Photonic
Locked Signal
accurate images/information because it captures and Laser Modulator Splitter
Processor
processes the vision/optical details of the targets/senses.
Of course, the accuracy of optical-based technology re-
mote sensing fully depends on the atmospheric/channel
conditions that limit the distance and performance qual- FIGURE 28. The schematic diagram of a photonic ADC.
ity. This section covers a review on the advancements
that have happened in the field of optical technology
for remote sensing: lidar, optical interferometry tech-
RF In RF Out
niques, holography techniques, IR/thermal imaging, and
aerial photography.
Laser Modulator Photodiode
Fiber-Optic
LIDAR Cable
Lidar was initially developed during the 1970s. It is an
active remote sensing technique/system similar to radar, FIGURE 29. A schematic diagram for an RoF scheme.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 97

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 optical interferometry sensor is shown in Figure 30(b). region. This allows the collection of a large amount of
The interference that is formed at the photodiode supports data from all directions of the volume under consider-
formulation of the image of the targets. These optical in- ation, enabling a fine range and Doppler resolution. The
terferometric methods are heavily used for astronomical three main substages in the processing of holographic
remote sensing. In the space, specialized methods and data are the detection engine, peak processor, and plot
suitable instruments are developed, like closure-phase filter [147]. In the first stage of processing (detection en-
imaging with multitelescope arrays for the purpose of as- gine), raw data are processed from all the receiver optics
tronomical remote sensing. Remote sensing of stars and and a 4D data matrix, which contains the range, azi-
planets through such optical interferometric methods muth, elevation, and pulse data, is formed. These data
allows us to study properties like dimensions, distances, are then converted to the frequency domain and thresh-
shapes, surface structures, and so on. Today’s optical in- old to identify the initial detections. The next stage,
terferometry-based remote sensing systems use several peak processing, refines detection by filtering based on
advanced photonics components to perform optical func- amplitude, Doppler, and so forth. In the final stage (plot
tions/processing, like beam transportation, beam combi- filter), targets are tracked using probability hypothesis
nation, optical modulation, polarization control, spatial density and reported to the target plot. Today, this type
filtering, and so forth [17]. of holography-based target sensing is used for noncoop-
erative sensing applications.
HOLOGRAPHY-BASED TARGET SENSING
Holographic radar is a variation of conventional radar, IR/THERMAL IMAGING
which uses a technology similar to holography for pro- IR/thermal imaging-based remote sensing is a technolo-
cessing the received optical signal and detecting a target. gy that uses the IR radiation of the target to create its op-
The working principle of holography-based remote sens- tical image at the IR region. A thermal IR imaging system
ing is shown in Figure 30(c). Wavefront changes carry requires unique cameras called IR cameras. The imaging
the structure of the targets, and they can be formed on a sensors in these cameras are sensitive to the wavelengths
film by comparing/combining them with the reference inside the IR location of the EM spectrum. IR radiation
plane wave, i.e., reference wave. If a particular volume is is not visible to the human eye, however, it is detectable
to be surveyed/imaged, a broad-beam antenna is used by IR cameras, and the imaging is also known as micro-
to illuminate the entire volume. A 2D array of optical scopic imaging. All the objects and scenes above absolute
receivers is used to receive the signals from the entire zero emit IR radiation, so thermal imaging cameras can

Reference Mirror
Lidar Transceiver
Light Beam
Light Transmitter Source Splitter Target
Source
Target

Photo
Receiver
Detector
Photo Detector
(a) (b)

Beam Splitter Target

Coherent Illumination
Light Wave

Reflection Target
Wave Wave

Reference
Sensing
Pure Reflection Wave
Film
Mirror

(c) (d) (e)

FIGURE 30. Optical technologies for remote sensing. (a) Lidar, (b) optical interferometry-based sensing, (c) holography-based sensing, and
(d) aerial photography.

98 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 create pictures of objects, even in mild, ambient condi- mainly for several applications like Earth imaging, map-
tions. Thermal imaging cameras are commonly available ping, inventorying, planning, defense applications, and
in two different frequency stages: standard frequency, in so on [112].
which the camera can produce images at a rate of nine
frames per second (fps), and HF, in which the cameras CHARACTERISTICS COMPARISON OF TODAY’S
produce images at a rate of 30 fps or more. A photo show- REMOTE SENSING SYSTEMS/TECHNOLOGIES
ing the optical image formed by a thermal camera, and A characteristics comparison of all of today’s active RF
photo of an IR camera are shown in Figure 30(d). The (radar systems), passive RF (EW systems), optical (vi-
operating principle of IR is primarily based on thermal/ sion systems) remote sensing systems, and technologies
IR emission of the object, and the quantity of the IR ra- is given in Table 2. Remote sensing systems/techniques,
diation emitted depends on its temperature, i.e., warmer their typical operating frequencies, and their potential
objects emit more radiation than cooler ones. IR-based applications are reported therein as the results of a com-
thermal imaging cameras also can locate the DoA of the parative study. In addition to the operating frequencies/
IR radiation emitted by objects, converting them into RF bands given in Table 2, radars/EW systems are op-
images. The general procedure of IR imaging is 1) the erated in/for some other frequency ranges based on the
camera lens focuses the radiation of the IR emitted by civilian/militar y requirements/applications. Typical
targets onto the detector chip, the detector chip is craft- potential applications for which the respective radar is
ed from plenty of tiny pixels, and every bit of it is respon- developed and/or EW systems are deployed are given in
sible for a particular range of the IR wavelength as soon Table 2, but the radars/EW systems are not absolutely
as the IR radiation touches a pixel, an electrical signal is limited to only those applications. Again, based on the
generated, the photograph is created by processing the requirements, the radar/EW systems are altered to per-
electrical signal produced by the pixel, and the amount form the other tasks as well. The main parameters, such
of radiation is nothing but the brightness or the intensity as type of radar, sensitivity/characteristics of EW sys-
of the pixel, thus, it is then displayed on the screen of tems, optical parameters, polarization, operating pow-
the camera. The main advantages with IR cameras are ers, and coverage ranges, wherever they are the major
IR-based thermal imaging supremacy over visible light limiting factor, are also given in Table 2 under the head-
imaging, they can be used in low-light or no-light condi- ing remarks. Based on a recent literature survey, mostly
tions as they detect IR radiation of all the objects above all of the remote sensing applications/technologies are
absolute zero, they can also measure the temperature of based on FMCW radar due to its various advantages, as
all objects, and they are nondestructive, i.e., they can detailed in the preceding sections. However, CW, pulsed,
examine objects without any physical contact or EM il- and pulsed-Doppler radars are also still being used for
lumination on objects. Further, they can be used to see some remote sensing applications. Microwave photon-
through smoke, fog, and other obstacles, basically, for ics is rapidly growing due to its inherent unique advan-
security and surveillance purposes. tages, and it appears that it will soon take over a major
part of today’s remote sensing technology. This means
AERIAL PHOTOGRAPHY that, except for the reception/transmission of microwave
Aerial photography is very simple, popular, and one of signals, transportation, ADCs, signal processing, micro-
the earliest methods of remote sensing [112]. Initially, it wave measurements, DACs, displays, and so forth will
used film-based cameras to capture images, but as tech- be only in the photonic domain, which will soon prove
nology developed, digital photography took over, provid- “gigahertz microwave signal processing by kilohertz
ing much better resolutions. Aerial photography initially electronics,” without any microwave downconversion
started with low-Earth-orbit satellites, and today, it is and high-frequency electrical ADCs.
popular even with drones. Photos of satellite- and drone-
based aerial photography are shown in Figure 30(e). In CONCLUSION
today’s advanced photography cameras, a resolution of In this article, we did a complete review of radar tech-
1:20,000, 0.2 m on the ground, is represented by one nology, right from the first attempt of radio detection,
pixel [112], [127]. Such improvements in resolution have to the current status of technologies and the future pos-
allowed satellites to have optical sensors/cameras, using sibilities. The article began with a history of radar sys-
them for satellite-based photography [112]. Further, de- tems, which highlighted all the major milestones in the
velopments in image processing have improved the ac- development of radar. Then, the basic working principles
curacy of target detection. The overall quality of aerial like the RCS and radar equation were covered, followed
photography has improved with contributions from tech- by the different ways to classify radars. Then, the article
nologies like removing smear from images with forward discussed the different applications of radar, like auto-
motion compression, reduction of tangential image mo- mobile, biomedical, and weather. Defense applications,
tion by gyro-stabilized mounts, and computer-based ex- which include EW systems, ECMs, ECCMs, and missile
posure controls [127]. Aerial photography today is used guidance systems, were discussed in detail. Following

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 99

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 TABLE 2. A CHARACTERISTICS COMPARISON OF TODAY’S REMOTE SENSING SYSTEMS/TECHNOLOGIES.

SR. REMOTE SENSING OPERATING

NUMBER SYSTEM/TECHNIQUE FREQUENCY POTENTIAL APPLICATIONS REMARKS

REMOTE SENSING RADAR SYSTEMS

1 Automotive radar 24–29 GHz, Vehicle detection, collision avoidance, parking FMCW and multiple-input,
76–81 GHz assistance, adaptive cruise control, automatic multiple-output (MIMO); max
emergency breaking, and lane change assistance. range: 100 m.
2 Biomedical radar 3–12 GHz, Human vital sign monitoring, fall detection, vision Doppler and FMCW; max range:
80 GHz, 288 GHz sensor, brain stroke/tumor detection, contactless 10 m.
human body scanning, and human body 3D imaging.
3 Weather radar 3–10 GHz Precipitation measurement, detection of hail, Pulse/Doppler radar, 50–200-km
tornadoes, blizzards, floods, rainfall, dam-tilt, range, and single polarization.
and water level/velocity.
4 Air traffic measurement 1,030–1,090 MHz, Airport surveillance, air route surveillance, and FMCW and MIMO; max range:
radar 2,700–2,900 MHz low-range radio altimeter. 5,000 m.
5 SAR 300 MHz–110 GHz Study of topography, oceanography, glaciology, Side-looking, FMCW, SFCW, and
geology, forestry, volcanoes, and earthquake dual- or quad-polarized sensors.
monitoring.
6 ISAR 8–12 GHz Detection of moving targets like cars, ships, boats, FMCW and SFCW.
vessels, airplanes, or spacecraft.
7 GPR 0.5–2 GHz Study of soil characteristics, water level, mining SFCW and multimonostatic/
activities, and land mine remediation; analysis of multibistatic radar; max range:
ice caps; and subsurface scanning for underground 500 m.
utilities.
8 Oil spill detection 4–8 GHz Detection of oil spills and study of ocean surface Application of SAR and SFCW;
radar changes. max range: 500 km.
9 FOPEN radar 30 MHz–4 GHz Military and civilian uses, crop monitoring,
foliage-hidden target detection, canopy density
detection, camouflage detection, and creating digital
maps under trees. Horizontally polarized and
application of SAR: max range: 100 m.
MODERN STARING RADAR SYSTEMS
10 MIMO radar 5.2 GHz, 8.9 GHz, Automotive radars, air traffic management radar, Pulse, FMCW, electronically
77 GHz and multiple-target tracking/guidance radar. steered array, and reduced SNR.
11 SDR 0.5–6 GHz Military/defense wireless sensor networks, software- CW, pulse, and FMCW radars.
defined reconfigurable radar, customized waveform
design, and compact for aerial applications.
12 Cognitive radar 300 MHz–8 GHz Image classification, land cover mapping, object FMCW radar.
detection, change detection, hyperspectral and radar
data analysis, and data fusion.
13 Photonics radar 0.5–30 GHz Lightweight/EMI-free radar design, extreme FMCW radar.
instantaneous bandwidth coverage, millimeter-range
resolution, sUAV target detection, oblique classification,
and compact for drone-based surveillance.
14 Quantum radar 18.75 GHz Stealth aircraft detection, jamming filters, and target Single microwave photon
detection in high-noise environments. transmission/reception.
15 Noncooperative 15 GHz Identification of malfunctioning satellite or debris, Pulse and FMCW radars.
target-sensing radar missiles or enemy aircraft, spy birds/UAVs, and collision
detection and avoidance targets/objects.
16 IoR 5 MHz–130 GHz Radar-data combined signal processing Centralized radar signal processing,
and analysis. imaging, and analysis and AI
engine-supported classification,
tracking, and guidance.
ADVANCED EW SYSTEM
17 ECM 0.5–30 GHz, has EW systems and identification of enemy radar/RF Extract all types of radar waveforms
more than two emitter’s signatures. and −80-dBm power.
split bands
18 Electronic counter 0.5–30 GHz, has Identification of enemy radar/RF emitter’s signature. Extract all types of radar waveforms
countermeasures more than two and −90-dBm power.
split bands
19 Photonic EW system 0.5–30 GHz, Identification of enemy radar/RF emitter’s signature. Extract all types of radar waveforms
instantaneous and −120-dBm power.
bandwidth

(Continued)

100 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 TABLE 2. A CHARACTERISTICS COMPARISON OF TODAY’S REMOTE SENSING SYSTEMS/TECHNOLOGIES. (CONTINUED)

SR. REMOTE SENSING OPERATING

NUMBER SYSTEM/TECHNIQUE FREQUENCY POTENTIAL APPLICATIONS REMARKS
20 Photonic signal Up to 300 GHz Accurate processing of RF signals in the photonics Used for the measurement of amp.,
processing domain. freq., phase. polarization. DoA, time
stamp, Doppler, PRT, pulsewidth,
bandwidth, and sweep rate.
21 Photonic ADC Up to 300 GHz High-speed RF sampling and quantization in the Gigahertz RF signal processing by
optical domain without downconversion. megahertz electronics.
22 Radio over fiber Up to 60 GHz High, instantaneous bandwidth RF transportation Low-loss RF transportation by
over single-mode fiber. ­intensity, phase, IQ, and QAM,
OFDM [180] schemes.
OPTICAL TECHNOLOGY FOR REMOTE SENSING
23 Lidar 200–400 THz Mapping landforms, ocean beds, and disaster Optical carriers are used to collect
management; autonomous cars; and infrastructure target information.
mapping.
24 Optical interferometry 400–790 THz Astronomical observations, measurement of 3D imaging and interference of
microscopic displacements, and refractive index waves.
changes.
25 Holography-based 100–500 MHz Humanitarian demining, and construction Accurate optical imaging of the
remote sensing inspection; survey of historic architecture, target.
paleontology, and security screening.
26 IR/thermal imaging 300 GHz–430 THz Sees through smoke and fog and detects cracks Nine- or 30-fps cameras and
in building structures; surveillance and power line temperature-dependent radiations.
inspection.
27 Aerial photography 400–790 THz Earth imaging, mapping, and inventorying; planning and An 80–200-km range and 1:20,000
defense applications. resolution.

EMI: electromagnetic interference; MIMO: multiple-input, multiple-output; max: maximum; amp.: amplitude; freq.: frequency; QAM: quadrature amplitude modulation.

that, we explored up-and-coming technologies in this AUTHOR INFORMATION

field, like MIMO radar, SDR, cognitive radar, quantum Samedh Sachin Kari ([email protected]) is
radar, and photonic radar. The final section explored a pursuing his B.Tech degree in electronics and commu-
very modern, advanced, and niche area of photonic EW nication engineering (avionics) with the Department of
systems, which completed the review on radar technol- Space, the Indian Institute of Space Science and Technol-
ogy. Near-future remote sensing technology will contain ogy (IIST), Thiruvananthapuram 659947, India. He is
a maximum of photonics sections, rather than classical presently working toward his bachelor’s degree in elec-
electrical/electronics due to the inherent advantages of tronics and communication engineering with a special-
photonics systems. Further, future remote sensing will ization in avionics with the Department of Avionics, IIST,
be in 3D (range, coordinate, and Doppler) and 4D (azi- India, and working for the development of radar-based
muth, elevation angle, slant range, and Doppler) spatial advanced remote sensing systems.
domains. AI will become an essential technique used A Arockia Bazil Raj ([email protected]) received his
to detect and classify near-future remote sensing data/ B.E. degree in electronics and communication engineering
images. Soon, mostly all of the detection/classification from Bharathidasan University, Tiruchirappalli, India. He
algorithms will run in embedded/hardware platforms to received his M.E. degree in communication systems and his
get advanced remote sensing systems with real-time sig- Ph.D. degree in information and communication technol-
nal processing technologies. ogy, both from Anna University, Chennai, India. He was an
assistant professor with the Kings College of Engineering
ACKNOWLEDGMENT from 2002 to 2006. He is currently an associate professor
Samedh Sachin Kari thanks Dr. A Arockia Bazil Raj and Dr. with the Defence Institute of Advanced Technology, Pune
Balasubramanian K. for their insights and guidance while 411025, India. Based on his teaching and research exper-
writing this review article. Thanks also to Defence Institute tise, he has written three books published by De Gruyter,
of Advance Technology (DIAT) staff, particularly Mahesh GmbH; CRC Press, USA; and the Institution of Engineering
Lonkar, lab assistant at the Radar Systems Laboratory, for and Technology, U.K. His current research interests cover
his support, and Dr. Kuruvilla Joseph, dean at the Indian radar system design, radar signal processing, real-time re-
Institute of Space Science and Technology for supporting mote sensing, free-space optical communication, digital
Kari’s internship at the Radar Systems Laboratory, DIAT, signal processing, and radio-frequency photonics. He is a
under Dr. Raj. Senior Member of IEEE.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 101

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 Balasubramanian K. ([email protected]) received his [12] M. Kameshwara Rao and A. Bazil Raj,” Reduced radar cross-
B.Tech. degree in chemical technology. He is a profes- section target imaging system,” in Proc. Int. Conf. Syst., Comput.,
sor, dean (academics), and head of the Department of Automat. Netw. (ICSCAN), Puducherry, India, 2021, pp. 1–6,
­Metallurgical of Materials Engineering, Ministry of De- doi: 10.1109/ICSCAN53069.2021.9526517.
fence, Defence Institute of Advanced Technology, Pune [13] C. Waldschmidt, P. Hügler, and M. Geiger, “D3.1 – Radar as an
411025, India. He is highly acclaimed for his contribu- emerging and growing technology for industrial applications:
tions to polymer processing and fabrication for various ap- A short overview,” in Proc. Sensor, 2017, pp. 460–465, doi:
plications, like smart textiles, superhydrophobic coatings, 10.5162/sensor2017/d3.1.
wastewater treatments, and polymer nanocomposites. Pri- [14] B. I. Bakare, M. U. Ajaegbu, and V. E. Idigo, “A comprehensive
or to his time at the Ministry of Defence, he worked as a review of radar system technology,” J. Electron. Commun. Eng.
research manager (technology), at the Materials Research Res., vol. 8, no. 8, pp. 16–23, 2022.
Institute (U.K.) for 10 years and transferred roughly eight [15] S. Latthe et al., “A review on ISAR imaging techniques for low
technologies to medium enterprises and small and medi- RCS targets,” Int. J. Eng. Res. Rev., vol. 10, no. 4, pp. 15–27, Oct./
um enterprises after receiving his Ph.D. degree in polymer Dec. 2022, doi: 10.5281/zenodo.7486091.
technology from the Institute of Polymer Technology and [16] E. D. Mantiply, K. R. Pohl, S. W. Poppell, and J. A. Murphy,
Materials Engineering, Loughborough University, U.K. “Summary of measured radiofrequency electric and mag-
netic fields (10 KHz to 30 GHz) in the general and work en-
REFERENCES vironment,” Bioelectromagnetics, vol. 18, no. 8, pp. 563–577,
[1] P. Amad, “From God’s-Eye to camera-eye: Aerial Photog- 1997, doi: 10.1002/(SICI)1521-186X(1997)18:8<563::AID-
raphy’s post-humanist and Neo-humanist visions of the BEM5>3.0.CO;2-0.
world,” Hist. Photogr., vol. 36, no. 1, pp. 66–86, Feb. 2012, doi: [17] R. Bamler and P. Hartl, “Synthetic aperture radar interfer-
10.1080/03087298.2012.632567. ometry,” Inverse Problems, vol. 14, no. 4, pp. 1–54, 1998, doi:
[2] M. Rast and T. H. Painter, “Earth observation imaging spectros- 10.1088/0266-5611/14/4/001.
copy for terrestrial systems: An overview of its history, tech- [18] IEEE Standard Letter Designations for Radar-Frequency Bands,
niques, and applications of its missions,” Surv. Geophys., vol. 40, IEEE Standard 521-2002 (Revision of IEEE Std 521-1984).
no. 3, pp. 303–331, 2019, doi: 10.1007/s10712-019-09517-z. [19] N. Ismail, T. Surya Gunawan, S. Kartika S, T. Praludi, and E.
[3] B. Kendal, “An overview of the development and introduction A. Z. Hamidi, “Design of microstrip hairpin bandpass filter
of ground radar to 1945,” J. Navigation, vol. 56, no. 3, pp. 343– for 2.9 GHz – 3.1 GHz S-band radar with defected ground
352, 2003, doi: 10.1017/s0373463303002406. structure,” Malaysian J. Fundam. Appl. Sci., vol. 14, no. 4,
[4] S. W. Lasswell, “History of SAR at Lockheed Martin (previously pp. 448–455, 2019, doi: 10.11113/mjfas.v14n4.1073.
Goodyear Aerospace),” in Proc. Radar Sensor Technol. IX, 2005, [20] A. Singh, V. Shah, and A. Sarnaik, “Moving target indication ra-
pp. 1–12, doi: 10.1117/12.603927. dar,” Int. J. Student Res. Technol. Manage., vol. 1, no. 1, pp. 27–38,
[5] B. Kaushik, D. Nance, and K. Ahuja, “A review of the role of Sep. 2015.
acoustic sensors in the Modern Battlefield,” in Proc. 11th AIAA/ [21] J.-M. Munoz-Ferreras, Z. Peng, R. Gomez-Garcia, and C. Li,
CEAS Aeroacoust. Conf., 2005, p. 2997, doi: 10.2514/6.2005-2997. “Review on advanced short-range multimode continuous-
[6] M. Guarnieri, “The early history of radar [historical],” IEEE wave radar architectures for healthcare applications,” IEEE J.
Instrum. Meas. Mag., vol. 4, no. 3, pp. 36–42, Sep. 2010, doi: Electromagn., RF, Microw. Med. Biol., vol. 1, no. 1, pp. 14–25, Jun.
10.1109/mie.2010.937936. 2017, doi: 10.1109/jerm.2017.2735241.
[7] S. Bennett, “History of automatic control to 1960: An over- [22] A. A. Raj et al., “A review–Unguided optical communications:
view,” IFAC Proc. Volumes, vol. 29, no. 1, pp. 3008–3013, 1996, Developments, technology evolution, and challenges,” Elec-
doi: 10.1016/s1474-6670(17)58135-5. tronics, vol. 12, no. 8, 2023, Art. no. 1922, doi: 10.3390/elec-
[8] S. M. N. Sakib, “LiDAR technology – An overview,” IUP J. Elect. tronics12081922.
Electron. Eng., vol. 15, no. 1, pp. 36–57, Jan. 2022. [23] M. Albéri et al., “Accuracy of flight altitude measured with
[9] D. Jayakumar, H. C. Kumawat, and A. A. B. Raj, “A new algo- low-cost GNSS, radar and barometer sensors: Implications for
rithm for automated LSS targets’ activities measurement based airborne radiometric surveys,” Sensors, vol. 17, no. 8, 2017, Art.
on a 24GHz CW RF sensor,” in Proc. Int. Conf. Advancement no. 1889, doi: 10.3390/s17081889.
Technol. (ICONAT), Goa, India, 2023, pp. 1–6, doi: 10.1109/ [24] S. Y. Matrosov, “Modeling backscatter properties of snowfall at
ICONAT57137.2023.10080042. millimeter wavelengths,” J. Atmospheric Sci., vol. 64, no. 5, pp.
[10] C. Uluisik, G. Cakir, M. Cakir, and L. Sevgi, “Radar cross sec- 1727–1736, 2007, doi: 10.1175/jas3904.1.
tion (RCS) modeling and simulation, part 1: A tutorial review [25] J. Xu, S. Zhu, and G. Liao, “Range ambiguous clutter suppres-
of definitions, strategies, and canonical examples,” IEEE Anten- sion for airborne FDA-STAP radar,” IEEE J. Sel. Topics Signal
nas Propag. Mag., vol. 50, no. 1, pp. 115–126, Feb. 2008, doi: Process., vol. 9, no. 8, pp. 1620–1631, Dec. 2015, doi: 10.1109/
10.1109/map.2008.4494511. jstsp.2015.2465353.
[11] T. E. Tice, “An overview of radar cross section measurement [26] R. S. Vickers, “Design and applications of airborne radars in
techniques,” in Proc. 6th IEEE Conf. Rec., Instrum. Meas. Technol. the VHF/UHF band,” IEEE Aerosp. Electron. Syst. Mag., vol. 17,
Conf., 1989, pp. 344–346, doi: 10.1109/imtc.1989.36882. no. 6, pp. 26–29, Jun. 2002, doi: 10.1109/MAES.2002.1010118.

102 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 [27] A. Battaglia et al., “Spaceborne cloud and precipitation ra- [43] R. Weinstein, “RFID: A technical overview and its application
dars: Status, challenges, and ways forward,” Rev. Geophys., to the enterprise,” IT Prof., vol. 7, no. 3, pp. 27–33, May/Jun.
vol. 58, no. 3, Sep. 2020, Art. no. e2019RG000686, doi: 2005, doi: 10.1109/MITP.2005.69.
10.1029/2019rg000686. [44] S. R. Rengarajan, “Advances in slotted waveguide array an-
[28] A. Herique, W. Kofman, P. Bauer, F. Remy, and L. Phalip- tenna technology,” in Proc. IEEE Int. Conf. Microw., Antennas,
pou, “A spaceborne ground penetrating radar: MIMOSA,” in Commun. Electron. Syst. (COMCAS), 2019, pp. 1–5, doi: 10.1109/
Proc. IEEE Int. Geosci. Remote Sens. Symp. (IGARSS), Hamburg, comcas44984.2019.8958050.
Germany, 1999, vol. 1, pp. 473–475, doi: 10.1109/IGARSS. [45] Y. E. Yamac and A. Kizilay, “A waveguide slot array antenna
1999.773537. design for X-band radar,” in Proc. 9th Int. Kharkiv Symp. Phys.
[29] K. Sankaran, “Spaceborne radar remote sensing of ocean sur- Eng. Microw., Millimeter Submillimeter Waves (MSMW), 2016,
faces: Electromagnetic modelling and applications,” J. Electro- pp. 1–5, doi: 10.1109/msmw.2016.7538075.
magn. Waves Appl., vol. 34, no. 1, pp. 1–34, 2019, doi: 10.1080/ [46] P. Delos, B. Broughton, and J. Kraft, “Phased array antenna
09205071.2019.1696237. patterns—Part 1: Linear array beam characteristics and array
[30] S. Zainuddin et al., “Maritime radar: A review on techniques factor,” Analog Dialog, vol. 54, no. 2, pp. 1–8, 2020.
for small vessels detection,” Int. J. Elect. Electron. Syst. Res., vol. [47] C. M. Seong and D. C. Park, “Design of cavity-backed spiral
14, pp. 30–45, Jun. 2019. antennas,” in Proc. 5th Global Symp. Millimeter-Waves, 2012,
[31] P. Braca, S. Maresca, R. Grasso, K. Bryan, and J. Horstmann, pp. 186–190, doi: 10.1109/gsmm.2012.6314032.
“Maritime surveillance with multiple over-the-horizon HFSW [48] D. Gao, “A design of lens antenna with high gain,” in Proc.
radars: An overview of recent experimentation,” IEEE Aerosp. MATEC Web Conf., 2nd Int. Conf. Comput. Sci. Commun. Netw.
Electron. Syst. Mag., vol. 30, no. 12, pp. 4–18, Dec. 2015, doi: Secur. (CSCNS), 2021, vol. 336, pp. 186–190, doi: 10.1051/
10.1109/maes.2015.150004. matecconf/202133601005.
[32] S. J. Anderson, F. Berizzi, and C. Bianchi, “Generation and [49] M. L. Van Blaricum, “A brief history of photonic antenna recon-
assimilation of propagation advice for HF skywave radar figuration,” in Proc. Int. Topical Meeting Microw. Photon. (MWP),
­systems,” in Proc. IEEE Radar Conf., Rome, Italy, 2008, pp. 1–6, 2000, pp. 1–5, doi: 10.1109/mwp.2000.889772.
doi: 10.1109/RADAR.2008.4720846. [50] T. Anderson, D. Melazzi, and V. Lancellotti, “An overview of
[33] J. Krolik, V. Mecca, O. Kazanci, and I. Bilik, “Multipath spread- experimental and numerical results on the performance of
Doppler clutter mitigation for over-the-horizon radar,” in Proc. plasma antennas arrays,” in Proc. 9th Eur. Conf. Antennas Propag.
IEEE Radar Conf., Rome, Italy, 2008, pp. 1–5, doi: 10.1109/ (EuCAP), 2015, pp. 1–4.
RADAR.2008.4720844. [51] K. Sushmaja, “Implementation of binary shift keying techniques,”
[34] J. D. Kraus, “Heinrich Hertz-theorist and experimenter,” IEEE Int. J. Eng. Trends Technol., vol. 4, no. 6, pp. 2581–2583, 2013.
Trans. Microw. Theory Techn., vol. 36, no. 5, pp. 824–829, May [52] P. Asuzu and C. Thompson, “A more exact linear FMCW ra-
1988, doi: 10.1109/22.3601. dar signal model for simultaneous range-velocity estimation,”
[35] A. Khan, M. Riaz, and A. Bilal, “Various types of antenna with in Proc. IEEE Radar Conf. (RadarConf), 2018, pp. 7–11, doi:
respect to their applications: A review,” Int. J. Multidisciplinary 10.1109/radar.2018.8378521.
Sci. Eng., vol. 7, no. 3, pp. 1–8, 2016. [53] M. Mercuri, D. Schreurs, and P. Leroux, “SFCW microwave
[36] S. Puri, K. Kaur, and N. Kumar, “A review of antennas for wire- radar for in-door fall detection,” in Proc. IEEE Topical Conf.
less communication devices,” Int. J. Electron. Elect. Eng., vol. 2, Biomed. Wireless Technol., Netw., and Sens. Syst. (BioWireleSS),
no. 3, pp. 199–201, 2014, doi: 10.12720/ijeee.2.3.199-201. 2012, pp. 53–56, doi: 10.1109/biowireless.2012.6172729.
[37] N. P. Agrawall, G. Kumar, and K. P. Ray, “Wide-band planar [54] T. Hwang, C. Yang, G. Wu, S. Li, and G. Ye Li, “OFDM and
monopole antennas,” IEEE Trans. Antennas Propag., vol. 46, its wireless applications: A survey,” IEEE Trans. Veh. Tech-
no. 2, pp. 294–295, Feb. 1998, doi: 10.1109/8.660976. nol., vol. 58, no. 4, pp. 1673–1694, May 2009, doi: 10.1109/
[38] C. Granet, “Designing axially symmetric Cassegrain or Grego- tvt.2008.2004555.
rian dual-reflector antennas from combinations of prescribed [55] C. S. Rawat, D. Balwani, D. Bedarkar, J. Lotwani, and H. K.
geometric parameters,” IEEE Antennas Propag. Mag., vol. 40, Saini, “Implementation of barker code and linear frequency
no. 2, pp. 76–82, Apr. 1998, doi: 10.1109/74.683545. modulation pulse compression techniques in Matlab,” Int. J.
[39] P. Dhande, “Antennas and its applications,” Sci. Spectr., vol. 2, Emerg. Technol. Adv. Eng., vol. 4, no. 4, pp. 105–111, Apr. 2014.
pp. 66–78, Mar. 2009. [56] G. Turin, “An introduction to matched filters,” IEEE Trans.
[40] J. D. Kraus, “The helical antenna,” Proc. IRE, vol. 37, no. 3, Inf. Theory, vol. 6, no. 3, pp. 311–329, Jun. 1960, doi: 10.1109/
pp. 263–272, Mar. 1949, doi: 10.1109/jrproc.1949.231279. tit.1960.1057571.
[41] R. Bawer and J. Wolfe, “The spiral antenna,” in Proc. IRE [57] H. A. Krichene et al., “Compressive sensing and stretch pro-
Int. Conv. Rec., 1960, pp. 84–95, doi: 10.1109/irecon.1960. cessing,” in Proc. IEEE RadarCon (RADAR), 2011, pp. 362–367,
1150893. doi: 10.1109/radar.2011.5960560.
[42] M. U. Khan, M. S. Sharawi, and R. Mittra, “Microstrip patch [58] K. Farnane, K. Minaoui, A. Rouijel, and D. Aboutajdine, “Anal-
antenna miniaturisation techniques: A review,” IET Microw., ysis of the ambiguity function for phase-coded waveforms,”
Antennas Propag., vol. 9, no. 9, pp. 913–922, 2015, doi: 10.1049/ in Proc. IEEE/ACS 12th Int. Conf. Comput. Syst. Appl. (AICCSA),
iet-map.2014.0602. 2015, pp. 1–4, doi: 10.1109/aiccsa.2015.7507195.

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 103

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 [59] M. K. Kıymık, İ. Güler, A. Dizibüyük, and M. Akın, “Com- sification using the range-/micro-Doppler signatures,” IEEE
parison of STFT and wavelet transform methods in deter- Trans. Radar Syst., vol. 1, pp. 623–631, 2023, doi: 10.1109/
mining epileptic seizure activity in EEG signals for real-time TRS.2023.3322607.
­application,” Comput. Biol. Med., vol. 35, no. 7, pp. 603–616, [74] H. C. Kumawat, M. Chakraborty, A. A. Bazil Raj, and S. V.
Oct. 2005, doi: 10.1016/j.compbiomed.2004.05.001. Dhavale, “DIAT-nSAT: Small aerial targets’ micro-Doppler
[60] X. Wu and T. Liu, “Spectral decomposition of seismic data signatures and their classification using CNN,” IEEE Geos-
with reassigned smoothed pseudo Wigner–Ville distribution,” ci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022, doi: 10.1109/
J. Appl. Geophys., vol. 68, no. 3, pp. 386–393, Jul. 2009, doi: LGRS.2021.3102039.
10.1016/j.jappgeo.2009.03.004. [75] M. Rupniewski, G. Mazurek, J. Gambrych, M. Nalecz, and R.
[61] H. C. Kumawat, A. B. Raj A, and P. Samczyński, “Spectrum Karolewski, “A real-time embedded heterogeneous GPU/FPGA
localization and Hough transform-based b tuning for LSS parallel system for radar signal processing,” in Proc. IEEE Conf.
targets’ accurate micro-Doppler imaging system,” IEEE Ubiquitous Intell. Comput., Adv. Trusted Comput., Scalable Com-
Trans. Instrum. Meas., vol. 71, pp. 1–11, 2022, doi: 10.1109/ put. Commun., Cloud Big Data Comput., Internet People, Smart
TIM.2022.3188540. World Congr. (UIC/ATC/ScalCom/CBDCom/IoP/SmartWorld),
[62] H. C. Kumawat and A. A. Raj, “SP-WVD with adaptive-filter- 2016, pp. 1189–1197, doi: 10.1109/uic-atc-scalcom-cbdcom-iop
bank-supported RF sensor for low RCS targets’ nonlinear -smartworld.2016.0182.
micro-Doppler signature/pattern imaging system,” Sensors, [76] R. A. A. Bazil, FPGA Based Embedded System Developer’s Guide.
vol. 22, no. 3, 2022, Art. no. 1186, doi: 10.3390/s22031186. Boca Raton, FL, USA: CRC Press, 2018.
[63] V. Winkler, “Range Doppler detection for automotive FMCW [77] M. L. Nascimento, “Brief history of X-ray tube patents,”
radars,” in Proc. Eur. Radar Conf., Munich, Germany, 2007, World Patent Inf., vol. 37, pp. 48–53, Jun. 2014, doi: 10.1016/
pp. 166–169, doi: 10.1109/EURAD.2007.4404963. j.wpi.2014.02.008.
[64] V. C. Chen and H. Ling, “Joint Time-frequency analysis for [78] G. Caryotakis, “The klystron: A microwave source of surprising
radar signal and image processing,” IEEE Signal Process. Mag., range and endurance,” Physics Plasmas, vol. 5, no. 5, pp. 1590–
vol. 16, no. 2, pp. 81–93, Mar. 1999, doi: 10.1109/79.752053. 1598, 1998, doi: 10.1063/1.872826.
[65] D. A. Robertson et al., “340-GHz 3D radar imaging test bed [79] Y. Blanchard, G. Galati, and P. van Genderen, “The cavity mag-
with 10-Hz frame rate,” in Proc. Int. Soc. Opt. Eng. (SPIE), 2012, netron: Not just a British invention [Historical Corner],” IEEE
pp. 41–51, doi: 10.1117/12.918581. Antennas Propag. Mag., vol. 55, no. 5, pp. 244–254, Oct. 2013,
[66] J. Zhang et al., “Application of deep generative networks doi: 10.1109/map.2013.6735528.
for SAR/ISAR: A review,” Artif. Intell. Rev., vol. 56, no. 10, [80] J. B. Gunn, “Microwave oscillations of current in III–V semi-
pp. 11,905–11,983, 2023, doi: 10.1007/s10462-023-10469-5. conductors,” Solid State Commun., vol. 1, no. 4, pp. 88–91, Sep.
[67] P. K. Eranti and B. D. Barkana, “An overview of direction-of- 1963, doi: 10.1016/0038-1098(63)90041-3.
arrival estimation methods using adaptive directional time- [81] D. M. Pozar, Microwave Engineering. Hoboken, NJ, USA: Wiley,
frequency distributions,” Electronics, vol. 11, no. 9, 2022, Art. 2021.
no. 1321, doi: 10.3390/electronics11091321. [82] A. Patyuchenko, “RF signal chain discourse—Part 2: Essential
[68] J. Ward, “Space-time adaptive processing for airborne radar,” building blocks,” Analog Dialog, vol. 55, no. 3, pp. 5–11, 2021.
in Proc. IEE Colloq. Space-Time Adaptive Process., 1998, pp. 1–6, [83] V. Reinhardt, “Direct digital synthesizers,” in Proc. 17th Annu.
doi: 10.1049/ic:19980240. Precise Time Time Interval Syst. Appl. Meeting, 1985, pp. 345–374.
[69] H. Lev-Ari and A. J. Devancy, “The time-reversal technique re- [84] N. Akhter, H. C. Kumawat, and A. Arockia Bazil Raj, “Develop-
interpreted: Subspace-based signal processing for multi-static ment of RF-photonic system for automatic targets’ nonlinear
target location,” in Proc. IEEE Sensor Array Multichannel Signal rotational/flapping/gliding signatures imaging applications,”
Process. Workshop (SAM), Cambridge, MA, USA, 2000, pp. 509– J. Circuits, Syst. Comput., vol. 32, no. 8, 2023, Art. no. 2350131,
513, doi: 10.1109/SAM.2000.878061. doi: 10.1142/s0218126623501311.
[70] Y. Jin, J. M. F. Moura, and N. O’Donoughue, “Time reversal in [85] V. Maheshwari, K. Sreenivasulu, M. Kumar, and K. Sharma,
multiple-input multiple-output radar,” IEEE J. Sel. Topics Sig- “Microwave photonic exciter unit for radar system,” 2019, arX-
nal Process., vol. 4, no. 1, pp. 210–225, Feb. 2010, doi: 10.1109/ iv:1910.08859.
JSTSP.2009.2038983. [86] D. Y. Lie, “‘RF-SoC’: Integration trends of on-chip CMOS pow-
[71] H. Deng, Z. Geng, and B. Himed, “Radar target detection using er amplifier: Benefits of external PA versus integrated PA for
target features and artificial intelligence,” in Proc. Int. Conf. Ra- portable wireless communications,” Int. J. Microw. Sci. Tech-
dar (RADAR), 2018, pp. 1–4, doi: 10.1109/radar.2018.8557249. nol., vol. 2010, pp. 1–7, Mar. 2010, doi: 10.1155/2010/380108.
[72] H. C. Kumawat, M. Chakraborty, and A. A. B. Raj, “DIAT-Rad- [87] M. Braun, J. Pendlum, and M. Ettus, “RFNoC™ — RF network-
SATNet—A novel lightweight DCNN architecture for micro- on-chip,” in Proc. 6th GNU Radio Conf., 2016, pp. 1–7.
Doppler-based small unmanned aerial vehicle (SUAV) targets’ [88] C. Waldschmidt, J. Hasch, and W. Menzel, “Automotive radar —
detection and classification,” IEEE Trans. Instrum. Meas., vol. From first efforts to future systems,” IEEE J. Microwaves, vol. 1,
71, pp. 1–11, 2022, doi: 10.1109/TIM.2022.3188050. no. 1, pp. 135–148, Jun. 2021, doi: 10.1109/jmw.2020.3033616.
[73] M. M. Hasan, M. Chakraborty, and A. A. B. Raj, “A hyper-pa- [89] S. M. Patole, M. Torlak, D. Wang, and M. Ali, “Automotive
rameters-tuned R-PCA+SVM technique for sUAV targets clas- radars: A review of signal processing techniques,” IEEE Signal

104 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 Process. Mag., vol. 34, no. 2, pp. 22–35, Mar. 2017, doi: 10.1109/ [105] Z. Sokol et al., “The role of weather radar in rainfall estimation
msp.2016.2628914. and its application in meteorological and hydrological model-
[90] M. Schneider, “Automotive radar – Status and trends,” in Proc. ling—A review,” Remote Sens., vol. 13, no. 3, 2021, Art. no. 351,
German Microw. Conf., Apr. 2015, pp. 144–147. doi: 10.3390/rs13030351.
[91] Y. Khraisat, “Simulation of the 24GHz short range, automotive [106] S. Trippe, “Polarization and polarimetry: A review,” J. Korean
radar,” Eur. J. Scientific Res., vol. 67, no. 1, pp. 75–82, 2011. Astronomical Soc., vol. 47, no. 1, pp. 15–39, Oct. 2014.
[92] H. Rohling and M.-M. Meinecke, “Waveform design principles [107] M. Kumjian, “Principles and applications of dual-polarization
for automotive radar systems,” in Proc. CIE Int. Conf. Radar weather radar. Part I: Description of the polarimetric radar
Proc., 2001, pp. 1–4, doi: 10.1109/icr.2001.984612. variables,” J. Oper. Meteorol., vol. 1, no. 19, pp. 226–242, 2013,
[93] S. Sun, A. P. Petropulu, and H. V. Poor, “MIMO radar for ad- doi: 10.15191/nwajom.2013.0119.
vanced driver-assistance systems and autonomous driving: Ad- [108] M. Skolnik, “An introduction to radar,” in Radar Handbook.
vantages and challenges,” IEEE Signal Process. Mag., vol. 37, no. New York, NY, USA: McGraw-Hill, 1962, pp. 1–21.
4, pp. 98–117, Jul. 2020, doi: 10.1109/msp.2020.2978507. [109] A. Pandey, M. Hasan, and A. A. B. Raj, “Space time adap-
[94] K. Tirumalesh and A. A. Bazil Raj, “Laboratory based auto- tive processing for high signal to clutter ratio target’s de-
motive radar for mobile targets ranging,” in Proc. Int. Conf. tection radar system,” in Proc. Int. Conf. Augmented Intell.
Syst., Comput., Automat. Netw. (ICSCAN), 2021, pp. 1–6, doi: Sustain. Syst. (ICAISS), 2022, pp. 1298–1302, doi: 10.1109/ic-
10.1109/icscan53069.2021.9526346. aiss55157.2022.10010908.
[95] R. Treuhaft, B. Law, and G. Asner, “Forest attributes from radar [110] F. J. Yanovsky, “Autonomous radio sensors for motion param-
interferometric structure and its fusion with optical remote eters,” in Aerospace Sensors, 1st ed. New York, NY, USA: Momen-
sensing,” BioScience, vol. 54, no. 6, pp. 561–571, Jun. 2004, doi: tum Press, 2013, pp. 89–136.
10.1641/0006-3568(2004)054[0561:FAFRIS]2.0.CO;2. [111] N. Gheitasi and F. Mazlumi, “A review on radar applications in
[96] C. Li, V. M. Lubecke, O. Boric-Lubecke, and J. Lin, “A review on civil aviation,” in Proc. Int. Conf. Iranian Aerosp. Soc., 2023, pp. 1–8.
recent advances in Doppler radar sensors for noncontact health- [112] J. L. Morgan, S. E. Gergel, and N. C. Coops, “Aerial photog-
care monitoring,” IEEE Trans. Microw. Theory Techn., vol. 61, no. 5, raphy: A rapidly evolving tool for ecological management,”
pp. 2046–2060, May 2013, doi: 10.1109/tmtt.2013.2256924. BioScience, vol. 60, no. 1, pp. 47–59, Jan. 2010, doi: 10.1525/
[97] J.-M. Munoz-Ferreras, J. Wang, Z. Peng, R. Gomez-Garcia, and bio.2010.60.1.9.
C. Li, “From Doppler to FMCW radars for non-contact vital- [113] J. A. Besada, I. Campaña, D. Carramiñana, L. Bergesio, and G.
sign monitoring,” in Proc. 2nd URSI Atlantic Radio Sci. Meet- de Miguel, “Review and simulation of counter-UAS sensors for
ing (at-RASC), 2018, pp. 1–4, doi: 10.23919/ursi-at-rasc.2018. unmanned traffic management,” Sensors, vol. 22, no. 1, 2021,
8471575. Art. no. 189, doi: 10.3390/s22010189.
[98] M. James, A. Vala, A. Patel, and H. Patel, “A brief review on bio- [114] G. Sharma and A. A. Raj, “OTFS based SAR with low complex-
medical application of radar,” Int. J. Eng. Sci. Invention, vol. 7, ity receiver,” IEEE Access, vol. 11, pp. 66,194–66,200, 2023, doi:
no. 4, pp. 66–72, Apr. 2018. 10.1109/access.2023.3290322.
[99] C. Li et al., “Overview of recent development on wireless sens- [115] G. Yadav, K. S. Rajpoot and A. A. Bazil Raj, “Synthetic aper-
ing circuits and systems for healthcare and biomedical applica- ture radar (SAR) image of small unmanned aerial vehicle
tions,” IEEE J. Emerg. Sel. Topics Circuits Syst., vol. 8, no. 2, pp. (sUAV),” in Proc. Int. Conf. Augmented Intell. Sustain. Syst. (IC-
165–177, Jun. 2018, doi: 10.1109/jetcas.2018.2822684. AISS), Trichy, India, 2022, pp. 1012–1018, doi: 10.1109/IC-
[100] D. Tang et al., “A DC-coupled high dynamic range biomedical AISS55157.2022.10011085.
radar sensor with fast-settling analog DC offset cancelation,” [116] A. Tsokas, M. Rysz, P. M. Pardalos, and K. Dipple, “SAR data ap-
IEEE Trans. Instrum. Meas., vol. 68, no. 5, pp. 1441–1450, May plications in earth observation: An overview,” Expert Syst. Appl.,
2019, doi: 10.1109/tim.2018.2888917. vol. 205, 2022, Art. no. 117342, doi: 10.1016/j.eswa.2022.117342.
[101] K. Mostov, E. Liptsen, and R. Boutchko, “Medical applications [117] G. Sharma and A. A. Bazil Raj, “Low complexity interference
of shortwave FM radar: Remote monitoring of cardiac and re- mitigation technique in IRCI-free SAR imaging algorithms,”
spiratory motion,” Med. Phys., vol. 37, no. 3, pp. 1332–1338, IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022, doi:
Mar. 2010, doi: 10.1118/1.3267038. 10.1109/LGRS.2022.3171340.
[102] M. A. Aldhaeebi et al., “Review of microwaves techniques for [118] F. Meyer, “Spaceborne synthetic aperture radar: Principles,
breast cancer detection,” Sensors, vol. 20, no. 8, 2020, Art. no. data access, and basic processing techniques,” in Synthetic
2390, doi: 10.3390/s20082390. Aperture Radar (SAR) Handbook: Comprehensive Methodologies
[103] E. Saltikoff et al., “An overview of using weather radar for cli- for Forest Monitoring and Biomass Estimation, SERVIR Global
matological studies: Successes, challenges, and potential,” Bull. Science Coordination Office, National Space Science and
Amer. Meterol. Soc., vol. 100, no. 9, pp. 1739–1752, 2019, doi: Technology Center, Huntsville, AL, USA, 2019, pp. 21–64.
10.1175/bams-d-18-0166.1. [119] G.-C. Sun et al., “Spaceborne synthetic aperture radar imaging
[104] M. Yeary, B. Cheong, J. Kurdzo, T. Yu, and R. Palmer, “A brief algorithms: An overview,” IEEE Geosci. Remote Sens. Mag., vol. 10,
overview of weather radar technologies and Instrumentation,” no. 1, pp. 161–184, Mar. 2022, doi: 10.1109/mgrs.2021.3097894.
IEEE Instrum. Meas. Mag., vol. 17, no. 5, pp. 10–15, Oct. 2014, [120] Z. Gao, C. Wei, C. Yang, Y. Xie, and H. Chen, “Parallel pro-
doi: 10.1109/mim.2014.6912194. cessing of sliding spotlight mode SAR imaging based on GPU,”

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 105

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 J. Eng., vol. 2019, no. 21, pp. 7607–7611, Nov. 2019, doi: via communication channels with bit-rate constraints,” Auto-
10.1049/joe.2019.0429. matica, vol. 42, no. 5, pp. 763–769, May 2006, doi: 10.1016/j.
[121] P. Berens, “Introduction to synthetic aperture radar (SAR),” in automatica.2005.12.024.
Advanced Radar Signal and Data Processing, Neuilly-sur-Seine, [135] B. A. White and A. Tsourdos, “Modern missile guidance design:
France: Nato Research and Technology Organization, 2006, An overview,” IFAC Proc. Volumes, vol. 34, no. 15, pp. 431–436,
pp. 3-1–3-14. Sep. 2001, doi: 10.1016/s1474-6670(17)40765-8.
[122] E. Giusti and M. Martorella, “Range Doppler and image au- [136] F. W. Nesline and P. Zarchan, “A new look at classical vs mod-
tofocusing for FMCW inverse synthetic aperture radar,” IEEE ern homing missile guidance,” J. Guid. Control, vol. 4, no. 1,
Trans. Aerosp. Electron. Syst., vol. 47, no. 4, pp. 2807–2823, Oct. pp. 78–85, 1981, doi: 10.2514/3.56054.
2011, doi: 10.1109/taes.2011.6034666. [137] G. El-Sheikh, “Performance evaluation of homing guidance
[123] R. Vehmas and N. Neuberger, “Inverse synthetic aperture radar methods,” in Proc. 14th Int. Conf. Aerosp. Sci. Aviation Technol.,
imaging: A historical perspective and state-of-the-art survey,” 2011, vol. 14, pp. 1–27, doi: 10.21608/asat.2011.23215.
IEEE Access, vol. 9, pp. 113,917–113,943, 2021, doi: 10.1109/ac- [138] E. Fishler et al., “MIMO radar: An idea whose time has come,”
cess.2021.3104799. in Proc. IEEE Radar Conf., 2004, pp. 71–78, doi: 10.1109/
[124] D. Comite, F. Ahmad, M. G. Amin, and T. Dogaru, “Forward- nrc.2004.1316398.
looking ground-penetrating radar: Subsurface target imaging and [139] A. R. Deepshikha and A. A. Bazil Raj, “Signal processing for
detection: A review,” IEEE Geosci. Remote Sens. Mag., vol. 9, no. 4, digital beamforming on transmit in MIMO radar,” in Proc. 2nd
pp. 173–190, Dec. 2021, doi: 10.1109/MGRS.2020.3048368. Int. Conf. Inventive Res. Comput. Appl. (ICIRCA), 2020, pp. 1106–
[125] I. Catapano, G. Gennarelli, G. Ludeno, C. Noviello, G. Es- 1111, doi: 10.1109/icirca48905.2020.9183290.
posito, and F. Soldovieri, “Contactless ground penetrating [140] A. Patyuchenko, T. Rommel, P. Laskowski, M. Younis, and G.
radar imaging: State of the art, challenges, and microwave Krieger, “Digital beam-forming reconfigurable radar system
tomography-based data processing,” IEEE Geosci. Remote Sens. demonstrator,” in Proc. IEEE Int. Geosci. Remote Sens. Symp.,
Mag., vol. 10, no. 1, pp. 251–273, Mar. 2022, doi: 10.1109/ 2012, pp. 1541–1544, doi: 10.1109/igarss.2012.6351103.
MGRS.2021.3082170. [141] M. A. Jensen and J. W. Wallace, “A review of antennas and
[126] S. Shahib, W. Al-Nuaimy, and A. Eriksen, “Image processing propagation for MIMO wireless communications,” IEEE Trans.
and neural network techniques for automatic detection and in- Antennas Propag., vol. 52, no. 11, pp. 2810–2824, Nov. 2004,
terpretation of ground penetrating radar data,” in Proc. 6th Int. doi: 10.1109/tap.2004.835272.
Multi-Conf. Circuits, Syst., Commun. Comput., 2002, pp. 12–16. [142] Y. Li, “MIMO radar waveform design: An overview,” J. Beijing
[127] R. J. Hall, “The roles of aerial photographs in forestry remote Inst. Technol., vol. 30, no. 1, pp. 44–59, 2021, doi: 10.15918/j.
sensing image analysis,” in Remote Sensing of Forest Environ- jbit1004-0579.2021.002.
ments. Boston, MA, USA: Springer-Verlag, 2003, pp. 47–75. [143] T. Debatty, “Software defined radar a state of the art,” in Proc.
[128] M. Fingas and C. Brown, “Review of oil spill remote sensing,” 2nd Int. Workshop Cogn. Inf. Process., 2010, pp. 253–257, doi:
Mar. Pollut. Bull., vol. 83, no. 1, pp. 9–23, 2014, doi: 10.1016/j. 10.1109/cip.2010.5604241.
marpolbul.2014.03.059. [144] T. Ulversoy, “Software defined radio: Challenges and opportu-
[129] K. Topouzelis, “Oil spill detection by SAR images: Dark nities,” IEEE Commun. Surveys Tuts., vol. 12, no. 4, pp. 531–550,
formation detection, feature extraction and classification Fourth Quart. 2010, doi: 10.1109/surv.2010.032910.00019.
­algorithms,” Sensors, vol. 8, no. 10, pp. 6642–6659, 2008, doi: [145] M. N. O. Sadiku and C. M. Akujuobi, “Software-defined radio:
10.3390/s8106642. A brief overview,” IEEE Potentials, vol. 23, no. 4, pp. 14–15,
[130] A. H. Solberg, C. Brekke, and P. O. Husoy, “Oil spill detection Oct./Nov. 2004, doi: 10.1109/mp.2004.1343223.
in Radarsat and Envisat SAR images,” IEEE Trans. Geosci. Re- [146] F. Harris and R. W. Lowdermilk, “Software defined radio: Part
mote Sens., vol. 45, no. 3, pp. 746–755, Mar. 2007, doi: 10.1109/ 22 in a series of tutorials on instrumentation and measure-
tgrs.2006.887019. ment,” IEEE Instrum. Meas. Mag., vol. 13, no. 1, pp. 23–32, Feb.
[131] N. Bhavsar et al., “A comprehensive review on foliage penetra- 2010, doi: 10.1109/mim.2010.5399214.
tion radar systems,” Int. J. Eng. Res. Rev., vol. 10, no. 4, pp. 28– [147] M. Jahangir and C. Baker, “Persistence surveillance of diffi-
40, 2022, doi: 10.5281/zenodo.7486148. cult to detect micro-drones with L-Band 3-D holographic ra-
[132] J. Richards, “The use of multiple-polarization data in foliage dar™,” in Proc. CIE Int. Conf. Radar (RADAR), 2016, pp. 1–5, doi:
penetrating (FOPEN) synthetic aperture radar (SAR) applica- 10.1109/radar.2016.8059282.
tions,” US Department of Energy (US), Washington, DC, USA, [148] V. Bamane et al., “A review on today’s cognitive radar technol-
Tech. Rep. SAND2002-2239 TRN: US200223%%147, 2002, ogy,” Int. J. Eng. Res. Rev., vol. 10, no. 4, pp. 41–51, 2022, doi:
doi: 10.2172/801391. 10.5281/zenodo.7486241.
[133] M. R. Allen, J. M. Jauregui, and L. E. Hoff, “FOPEN-SAR de- [149] A. F. Martone, “Cognitive radar demystified,” URSI Radio Sci.
tection by direct use of simple scattering physics,” in Proc. Int. Bull., vol. 2014, no. 350, pp. 10–22, Sep. 2014, doi: 10.23919/
Radar Conf., Alexandria, VA, USA, 1995, pp. 152–157, doi: URSIRSB.2014.7910111.
10.1109/RADAR.1995.522537. [150] S. Z. Gurbuz et al., “An overview of cognitive radar: Past, pres-
[134] V. Malyavej, I. R. Manchester, and A. V. Savkin, “Precision ent, and future,” IEEE Aerosp. Electron. Syst. Mag., vol. 34, no. 12,
missile guidance using radar/multiple-video sensor fusion pp. 6–18, Dec. 2019, doi: 10.1109/maes.2019.2953762.

106 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE MARCH 2024

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.
 [151] U. Garg, A. A. Bazil Raj, and K. P. Ray, “Cognitive radar as- [167] F. A. Butt and M. Jalil, “An overview of electronic warfare in
sisted target tracking: A study,” in Proc. 3rd Int. Conf. Commun. radar systems,” in Proc. Int. Conf. Technol. Adv. Elect., Electron.
­Electron. Syst. (ICCES), Coimbatore, India, 2018, pp. 427–430, Comput. Eng. (TAEECE), 2013, pp. 213–217, doi: 10.1109/
doi: 10.1109/CESYS.2018.8723894. taeece.2013.6557273.
[152] M. Chakraborty, H. C. Kumawat, S. V. Dhavale, and A. B. Raj A, [168] A. Abdalla, H. Abdalla, M. Ramadan, S. Mohamed, and T.
“DIAT-RadHARNet: A lightweight DCNN for radar based classifi- Bin, “Overview of frequency diverse array in radar ECCM
cation of human suspicious activities,” IEEE Trans. Instrum. Meas., applications,” in Proc. Int. Conf. Commun., Control, Com-
vol. 71, pp. 1–10, 2022, doi: 10.1109/TIM.2022.3154832. put. Electron. Eng. (ICCCCEE), 2017, pp. 1–9, doi: 10.1109/
[153] K. S. Rajpoot and A. A. Raj, “Cognitive radar for target tracking,” in iccccee.2017.7867610.
Proc. Int. Conf. Trends Quantum Comput. Emerg. Bus. Technol. (TQCEBT), [169] M. E. Manka, “Microwave photonics electronic warfare tech-
2022, pp. 1–5, doi: 10.1109/tqcebt54229.2022.10041644. nologies for Australian defence,” in Proc. IEEE Avionics, Fiber-
[154] T. Borowski, M. Pasternak, and J. Pietrasinski, “The photonic Opt. Phototonics Technol. Conf., San Antonio, TX, USA, 2009,
radar: The situation today and the prospects for the future,” pp. 1–2, doi: 10.1109/AVFOP.2009.5342738.
in Proc. Radioelectronic Syst. Conf., 2020, pp. 175–185, doi: [170] M. E. Manka, “Microwave photonics for electronic war-
10.1117/12.2565721. fare applications,” in Proc. Int. Topical Meeting Microw. Pho-
[155] S. De and A. A. Bazil Raj, “Photonics radar modeling and simu- ton. Jointly Held Asia-Pacific Microw. Photon. Conf., Gold
lation in opti-system platform for high resolution target detec- Coast, QLD, Australia, 2008, pp. 275–278, doi: 10.1109/
tion,” in Proc. Int. Conf. Smart Gener. Comput., Commun. Netw. MWP.2008.4666690.
(SMART GENCON), Pune, India, 2021, pp. 1–7, doi: 10.1109/ [171] R. A. Minasian, “Photonic signal processing of microwave
SMARTGENCON51891.2021.9645780. signals,” IEEE Trans. Microw. Theory Techn., vol. 54, no. 2, pp.
[156] S. De and A. A. Bazil Raj, “A survey on photonics technologies 832–846, Feb. 2006, doi: 10.1109/TMTT.2005.863060.
for radar applications,” J. Opt., vol. 52, no. 1, pp. 90–119, Mar. [172] J. Capmany, J. Mora, I. Gasulla, J. Sancho, J. Lloret, and S.
2023, doi: 10.1007/s12596-022-00897-x. Sales, “Microwave photonic signal processing,” J. Lightwave
[157] F. Zhang, Q. Guo, and S. Pan, “Photonics-based real-time ul- Technol., vol. 31, no. 4, pp. 571–586, Feb. 2013, doi: 10.1109/
tra-high-range-resolution radar with broadband signal genera- JLT.2012.2222348.
tion and processing,” Scientific Rep., vol. 7, no. 1, 2017, Art. no. [173] G. C. Valley, “Photonic analog-to-digital converters,” Opt. Ex-
13848, doi: 10.1038/s41598-017-14306-y. press, vol. 15, no. 5, pp. 1955–1982, 2007, doi: 10.1364/OE.15.
[158] N. Akhter, A. A. Raj, and K. Prabu, “Photonics radar based LSS 001955.
targets’ postures’ m-D and cadence frequency imaging using [174] Y. Han and B. Jalali, “Photonic time-stretched analog-to-dig-
empirical wavelet transform technique,” Imag. Sci. J., early ac- ital converter: Fundamental concepts and practical consider-
cess, 2023, doi: 10.1080/13682199.2023.2221052. ations,” J. Lightwave Technol., vol. 21, no. 12, pp. 3085–3103,
[159] P. Ghelfi et al., “A fully photonics-based coherent radar sys- Dec. 2003, doi: 10.1109/JLT.2003.821731.
tem,” Nature, vol. 507, no. 7492, pp. 341–345, Mar. 2014, doi: [175] S. R. Abdollahi, H. S. Al-Raweshidy, S. M. Fakhraie, and R.
10.1038/nature13078. Nilavalan, “Digital Radio over Fibre for Future Broadband
[160] R. Torromé, N. Bekhti-Winkel, and P. Knott, “Introduction to Wireless Access Network Solution,” in Proc. 6th Int. Conf. Wire-
quantum radar,” Jan. 2021, arXiv:2006.14238. less Mobile Commun., Valencia, Spain, 2010, pp. 504–508, doi:
[161] M. Frasca, A. Farina, and B. Balaji, “Foreword to the special is- 10.1109/ICWMC.2010.32.
sue on quantum radar,” IEEE Aerosp. Electron. Syst. Mag., vol. 35, [176] C. Lim, Y. Tian, C. Ranaweera, T. A. Nirmalathas, E. Wong,
no. 4, pp. 4–7, Apr. 2020, doi: 10.1109/maes.2020.2977851. and K.-L. Lee, “Evolution of radio-over-fiber technology,”
[162] M. Lanzagorta and J. Uhlmann, “Opportunities and challenges J. Lightwave Technol., vol. 37, no. 6, pp. 1647–1656, Mar. 2019,
of quantum radar,” IEEE Aerosp. Electron. Syst. Mag., vol. 35, no. doi: 10.1109/JLT.2018.2876722.
11, pp. 38–56, Nov. 2020, doi: 10.1109/maes.2020.3004053. [177] A. M. Zin, M. S. Bongsu, S. M. Idrus, and N. Zulkifli, “An
[163] Z. Jing, H. Pan, Y. Li, and P. Dong, “Introduction,” in Non- overview of radio-over-fiber network technology,” in Proc. Int.
Cooperative Target Tracking, Fusion and Control. Information Fusion Conf. Photon., Langkawi, Malaysia, 2010, pp. 1–3, doi: 10.1109/
and Data Science. Cham, Switzerland: Springer-Verlag, 2018, ICP.2010.5604429.
pp. 3–6. [178] S. J. Purkis and J. C. Brock, “LiDAR overview,” Coral Reef Remote
[164] S. K. Wong, “Non-cooperative target recognition in the fre- Sensing. Dordrecht, The Netherlands: Springer-Verlag, 2013,
quency domain,” IEE Proc. – Radar, Sonar Navigation, vol. 151, pp. 115–143.
no. 2, pp. 77–84, 2004, doi: 10.1049/ip-rsn:20040197. [179] R. Harrap and M. Lato, “An overview of LIDAR: Collection to
[165] Y. K. Kwag and C. H. Chung, “UAV based collision avoidance applications,” NGI Publication, vol. 2, pp. 1–9, Jan. 2010.
radar sensor,” in Proc. IEEE Int. Geosci. Remote Sens. Symp., 2007, [180] N. Yadav and A. Rockia Bazil Raj, “Orthogonal frequency
pp. 639–642, doi: 10.1109/igarss.2007.4422877. division multiplexing waveform based radar system,” in
[166] P. Sharma, K. K. Sarma, and N. E. Mastorakis, “Artificial in- Proc. Int. Conf. Syst. Comput., Automat. Netw. (ICSCAN), Pu-
telligence aided electronic warfare systems- recent trends ducherry, India, 2021, pp. 1–4, doi: 10.1109/ICSCAN53069.
and evolving applications,” IEEE Access, vol. 8, pp. 224,761– 2021.9526501.
224,780, 2020, doi: 10.1109/access.2020.3044453. GRS

MARCH 2024 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE 107

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA. Downloaded on August 01,2024 at 05:27:09 UTC from IEEE Xplore. Restrictions apply.