ISSN (Print) : 0974-6846

Indian Journal of Science and Technology, Vol 9(47), DOI: 10.17485/ijst/2016/v9i47/100967, December 2016 ISSN (Online) : 0974-5645

Use of Matched Filter in Direct Sequence

Spread Spectrum Techniques
Iffat Rehman Ansari1, Shahnawaz Uddin1 and Sameena Naaz2*
1
University Women’s Polytechnic, Aligarh Muslim University, Aligarh – 202002, Uttar Pradesh, India;
[email protected], [email protected]
2
Department of Computer Science and Engineering, Faculty of Engineering and Technology,
Near Batra Hospital, Hamdard Nagar, New Delhi – 110062, India;
[email protected]

Abstract
Objectives: The protection of signal against jamming is accomplished using Spread Spectrum (SS) techniques. In this research
paper, noise rejecting property of the matched filter and Binary Phase Shift Keying (BPSK) modulation is used in Direct Sequence
Spread Spectrum (DSSS) system for secure communication. Methods/Statistical Analysis: Here, DSSS system has been
developed and simulated using C-programming language on windows platform. In order to perform the simulation, different
parameters are initialized and message bits are generated by random function. With the help of binary message signal and PN
sequence, a SS signal is generated which is modulated using BPSK modulation technique at the transmitter side. In the channel,
noise/jamming signal is added to the transmitted signal. Then the received signal is demodulated using coherent BPSK receiver
(i.e., correlation receiver) and matched filtering is followed by despreading the demodulated signal at the receiver side. Finally,
the received binary message bits are detected by threshold detector and various parameters such as input SNR, output SNR,
probability of error are calculated. Findings: It has been found that matched filter is a linear filter which has a noise rejection
property (means jamming signals can be treated just like other random noise) and hence provides signal security. This particular
property of matched filter is used in DSSS communication. As far as matched filter is concerned, it cannot remove the total effect
of interference but still the waveforms can be detected in the presence of noise/jamming signal. So, a matched filter receiver
is supposed to be an ideal receiver for a signal which is corrupted by AWGN/jamming signal. Here, an average probability of
symbol error decreases because it maximizes the output signal to noise ratio. These parameters are supposed to be the most
important factors that determine the efficiency of performance of the system. Applications: DSSS systems using matched filter
are strong contender for mobile and personnel communications, due to their very large user handling capacity. They also have
better transmission quality and capability by incorporating wireless services such as data, short transaction messages, and
multimedia etc.

Keywords: Signal Security,Matched filter, DSSS, PN Sequence, SNR, Probability of Error.

1.  Introduction to unauthenticated receivers by making various trans-

formations. On the other hand, spread spectrum (SS)
There are various methods of message security for modulation is among the methods being actively used
instance, encryption, spread spectrum modulation etc. to provide protection to the transmitted signal against
In encryption, the presence of message is not concealed, intentional and unintentional interference. In SS modula-
rather the information content of the message is denied tion, the data to be sent occupies a bandwidth in larger

*Author for correspondence

 Use of Matched Filter in Direct Sequence Spread Spectrum Techniques

than the minimum bandwidth necessary to send it. As the the presence of noise. Forward Error Correction (FEC)6
modulated signal occupies a far wider bandwidth than is another mechanism used to improve the reliability of
the bandwidth required by the original information sig- digital communications over wireless channels in pres-
nal, it is called spread spectrum modulation technique. ence of interference, fading, and noise. The median filter
The spreading of spectrum is achieved before transmis- may also efficiently detect the original signal in presence
sion by the use of a code that does not depend on the data of bipolar impulse noise 7.
sequence. The receiver operates in synchronization with
the transmitter and uses the same code to de-spread the
received signal so that the original data may be recovered.
2.  An Overview of Matched Filter
Spread spectrum methods have been in use for anti- A lot of techniques have been developed and implemented
jamming and multipath rejection applications as well as for smoothing denoising filters. Spatial domain filters can
for accurate ranging and tracking. Legitimate user’s com- be used to reduce noise8. Some of these filters have been
munication can be effectively disrupted by jamming unless described in the references9-13. The two most promising
they are equipped with extremely sophisticated detection techniques developed in the last two decades are wave-
mechanisms and countermeasures. Although DSSS has let analysis and matched filter14-15. Matched filter gives
an anti- jamming capability, there also exists intelligent optimum signal to noise ratio (SNR) of the signal at the
jamming strategies that effectively attack a DSSS channel. receiver side16. Wavelet matched filters use wavelets along
There has been active research on developing effective with Fourier transform for recognition and classification
jammer strategies, e.g., studied channel-oblivious and of two dimensional objects. Phase-only Wavelet Matched
channel-aware jammers8. Spread spectrum systems are Filter has been discussed in the reference17.
designed to work against interference of finite power.
This interference could for example be due to alternate
2.1  Matched Filter Receiver
propagation paths of the signal itself, as in multipath fad-
ing, or other transmitters. In spread spectrum systems, A linear filter with impulse response, hj(t), with received
bandwidth is sacrificed for reliability and low levels of signal x(t) used as the filter input, the resulting filter out-
interference with other types of communication. put, yj(t), is defined by the convolution integral:
Direct sequence spread spectrum communication
utilizes the noise rejection property of the Matched fil- (1)
ter. Rejection of interference, resistance against jamming, Suppose, the impulse response is:
privacy of message and a number of other important
properties are provided by DSSS technique1. As an exam- (2)
ple we can specify that the wide band signal cannot be
Then, the resulting filter output is:
jammed by a jamming signal if the spread spectrum pat-
tern and algorithm for generating the spread sequence is
not known2. Specifically, with the interference rejection (3)
property, commercial wireless network standards such Sampling this output at time t=T, we get:
as IEEE 802.11 and IEEE 802.15.4 widely adopt DSSS, to
provide robust communications3-4.
Various demodulation/decoding techniques may be (4)
used for correctly reproducing the transmitted signal at Since ϕj(T) is zero outside the interval 0 ≤ t ≤ T, we finally get:
the receiver end in the presence of noise but the optimum
receiver for a signal corrupted by Additive White Gaussian
Noise (AWGN) is the Matched filter receiver5. Matched (5)
filter receivers have also been shown to be very robust
against many other forms of noise, and have thus been It is to be noted that yj(T) = xj , where xj is the jth cor-
employed in many more applications. Not all inference relator output produced by the received signal x(t)18. Thus
will be removed by the matched filter, it can be designed the detector part of the optimum receiver may also be
such that the original signals are correctly decoded in implemented as in Figure 1.

2 Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology
 Iffat Rehman Ansari, Shahnawaz Uddin and Sameena Naaz

Figure 2.  Model of direct sequence spread Spectrum PSK

Figure 1.  Detector part of matched filter receiver. System.

A filter is said to be matched to ϕj(t) if its impulse

3.1  PN (Pseudo-Noise) Sequences:
response is a delayed and time-reversed version of some sig-
nal ϕj(t) , as in Eq. (2). The optimum receiver based on the A coded sequence of 1’s and 0’s having certain auto-
detector of Figure 1 is said to be a matched filter receiver. For correlation properties is known as PN sequence.
a matched filter operating in real time to be physically real- In spread spectrum communication systems, peri-
izable, it must be causal. That is to say, its impulse response odic PN sequences are used. PN sequences may be
must be zero for negative time, as shown by: aperiodic and such sequences are known as Barker
sequences. In order to generate the maximum length
hj (t) = 0 ; t < 0 (6) PN sequence, linear feedback shift register may be
used. In a shift register having length m, m flip- flops
With hj(t) defined in terms of ϕj(t) as in Eq. (2), we see
are used and they are controlled by a single timing
that the causality condition is satisfied provided that the
clock and the state of each flip-flop is shifted to the
signal ϕj(t) is zero outside the interval 0 ≤ t ≤ T 19.
next one at each pulse of the clock. The feedback
function is obtained as shown in Figure 3.
3.  System Model
In order to transmit the signal, the spread spectrum is 1 2 m

carried out before phase modulation. For the purpose of

analysis, however, it will be convenient to interchange the
order of these two operation in the transmitter side. It is
permitted to do this because the spectrum spreading and
the phase shift keying are both linear operation.
Figure 2 illustrates the model of direct sequence
spread spectrum PSK system. In this model, it is assumed
that the effect of channel noise may be ignored because
the interference limits the performance. Therefore, the Figure 3.  Maximum length sequence generator.
output of the channel is obtained by multiplying the
Pseudo Noise sequence with the Phase Shift Keyed signal 4.  Simulation of Matched Filter
and then adding the interference to it. The received signal
is first multiplied by the PN sequence at the receiver, then
Receiver
it is fed to the coherent detector. The coherent detector
input consists of a PSK signal embedded in additive code
4.1 Use of M-Sequence in Direct Sequence
modulated interference. The modulated nature of the lat- Spread Spectrum:
ter component forces the interference signal (jammer) to 1. The polar form of the data sequence d(t) is assumed,
spread its spectrum, such that the detection of informa- i.e., the bits are represented by +1.
tion bit at the receiver output is afforded which results in 2. Then, the m-sequence produced by the feedback shift
increased reliability. register is basically the spreading PN-code.

Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology 3
 Use of Matched Filter in Direct Sequence Spread Spectrum Techniques

3. The error performance is simulated in the presence denoted by ±1 and feedback shift register of length m
of one sinusoidal interfering/jamming signal having =4 is used to produce the PN sequence. The RF carrier
frequency of 0.49 kHz, where message sequence is is modulated with the product of data sequence and PN
produced by random number generation. sequence using binary phase shift keying (BPSK) at the
4. The error performance is again simulated in the pres- transmitter side. An interfering/jamming signal is intro-
ence of other sinusoidal interfering/jamming signal duced in the channel. The BPSK demodulator is used to
with frequencies of 0.339 kHz, 0.99 kHz, and 1.34 kHz demodulate the received signal at the receiver side and
respectively. then the output of the demodulator is multiplied by a PN
5. The error performance having two sinusoidal jammers sequence to nullify the effect of PN-code spreading. The
of equal amplitude with the frequencies of 0.49 kHz coherent BPSK receiver (i.e., correlation receiver) is fol-
and 0.339 kHz is also simulated. lowed by a matched filter, the output of which is passed
6. Then, the (SNR)I and (SNR)O are calculated in each case. through a decision device where the original message sig-
nal is estimated.
4.2 Block Diagram for PN Sequence
Generation
The memory shift register of m=4 produces the
m-sequence having length of 15. In order to generate the
m-sequence, initial states are presumed to be 1111. This is
shown in Figure 4.

x1 Flip Flop x2 Flip Flop x3 Flip Flop x4

Flip Flop
1 2 3 4

Modulo - 2

Figure 5.  Direct sequence BPSK SS system.

Output
Sequence
2 -1
The implementation of the simulation has been
Figure 4.  Feedback shift register for the m-sequence of carried out in C language. In order to simulate a
period 15. direct sequence spread spectrum system, m- sequence
is used as the spreading code and the value of m=4.
The states of four flip flops are represented as x1, x2, x3 Here, two sinusoidal jamming signals are generated.
and x4 for m= 4. Input and output signal to noise ratio are calculated
The code generated is like this: 1, 1, 1, 1, -1, 1, -1, 1, 1, along with the error probability and same have been
-1, -1, 1, -1, -1, -1 displayed.
Here, m-sequence is suited to this task because they
have a broad spectrum since the runs of the same symbols
are distributed.
5.  Results and Discussion
4.3 Direct Sequence BPSK Spread Spectrum
Figure 6 shows the binary form of the m-sequence,
System its polar waveform representation c(t), and the auto-
Figure 5 depicts a direct sequence spread spectrum sys- correlation function. The auto-correlation function of
tem where an m-sequence is used as the spreading code m-sequence is of periodic type, binary valued and it
and BPSK as the modulation technique. The data bits are resembles like noise.

4 Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology
 Iffat Rehman Ansari, Shahnawaz Uddin and Sameena Naaz

On the other hand, if an alternating sequence of 1’s

and 0’s is added to the modulated m-sequence, this sort
of the interference is equivalent to jamming the DS/BPSK
signal with a coherent BPSK signal of the same carrier
frequency. This is much more serious form of jamming
because it can cause oscillations to appear in the matched
filter output.
Simulation results are given in Table 1 and Table 2.
Here, both input signal to noise ratio (SNR)I and output
signal to noise ratio (SNR)O are calculated and compared.
It can be seen from the tables that an increase in the out-
Figure 6. Polar waveform of m-sequence and put signal to noise ratio results in decrease of probability
autocorrelation function of symbol error.

Figure 7 illustrates the data sequence d(t): [1, 1, 0, 1, 0, 1, Table 1.  Simulation results for a single sinusoidal jammer
0, 0, 0, 1, 1, 1, 1, 0, 0], the modulated wave m(t), and the Frequency Amplitude Pe (SNR)I (dB) (SNR)O
output of the matched filter y(t). This output is same as (KHz) (volts) (dB)
the auto-correlation function of m- sequence. There are 0.49 8 0.2915 -15.05 -1.61
side lobes present in the filter output for an alternating
0.49 6 0.2061 -12.55 0.88
sequence of 1’s and 0’s, and these side lobes are still 4.8 dB
0.49 5 0.0988 -10.97 2.47
below the peaks at the appropriate sampling instant.
0.49 4.8 0.0473 -10.61 2.83
0.49 4.6 0 -10.24 3.20
0.339 4 0.2793 -9.03 -1.25
0.339 3 0.1715 -6.53 1.33
0.339 2.5 0.0515 -4.95 2.89
0.339 2.4 0 -4.59 3.23
0.339 2.3 0 -4.22 3.72
0.99 8 0 -15.05 40.34
1.34 18 0.0941 -22.10 2.62
1.34 17 0.0133 -21.59 3.10
1.34 16.5 0.0049 -21.34 3.36
Figure 7.  Modulated signal waveform and output 1.34 16.0 0 -21.07 3.63
waveform of matched filter

It is to be noted that if a DC interference is added Table 2.  Simulation results for two sinusoidal jammers
to the modulated wave m(t), then the modulated wave- Amplitude Amplitude Pe (SNR)I (SNR)O
form (in a bipolar format) is converted into a unipolar (in volts) of (in volts) of (dB) (dB)
format having amplitude 2. This sort of the interfer- Jammer at Jammer at
ence is the baseband equivalent of jamming a direct 0.45 KHz 0.339 KHz
sequence BPSK spread spectrum system by a single 2.0 2.0 0.0655 -6.02 3.77
tone jammer at different varying frequencies. This is 1.8 1.8 0.0314 -5.11 4.68
not a very effective jamming method, since its only 1.7 1.7 0.0151 -4.61 5.17
effect is to add a bias term in the matched filter out- 1.65 1.65 0.0062 -4.35 5.43
put. Further, the effect of this jammer can be entirely 1.62 1.62 0.0008 -4.19 5.59
removed by adaptively changing the threshold of the
1.60 1.60 0 -4.08 5.70
decision device.

Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology 5
 Use of Matched Filter in Direct Sequence Spread Spectrum Techniques

5.1 Probability of Error Vs (SNR)O for input signal and output signal have been compared and it
DSSS BPSK Signal is clear that an increase in the output signal to noise ratio
results in the decrease of average probability of symbol
As far as the coherent BPSK is concerned, the average error. However, this decrease in the probability of symbol
probability of symbol error is defined as: error (i.e., an increase in the reliability) is achieved at the
cost of system bandwidth.
Pe = 1/2 erfc (Eb/No)1/2 (7)

where, Eb = Transmitted Bit Energy 7.  References

No = Noise power of AWGN
The complementary error function is defined as: 1. Lee J S, Miller L E. CDMA Systems Engineering Handbook,
Norwood, MA: Artech House Inc., 1998.
(8) 2. Simon M, Omura J, Scholtz R, Levitt B. Spread Spectrum
Communications Handbook, New York, NY: McGraw-
and we know that, erfc(u) = Hill, 1994.
3. IEEE: Institute of Electrical and Electronics Engineers,
IEEE Standards for Local and Metropolitan Area
From the graph shown in Figure 8, it can be clearly seen Networks, Part 11: Wireless LAN Medium Access
that as the output signal to noise ratio (SNR)O increases, Control (MAC) and Physical Layer (PHY) Specifications,
the average probability of symbol error Pe decreases. This IEEE Standard 1997.
basically explains the nature of the curve. 4. IEEE: Institute of Electrical and Electronics Engineers,
IEEE Standard for Information Technology: Local and
Y Metropolitan Area Networks, Specific Requirement,
13 Part 15.4: Wireless Medium Access Control (MAC)
12 and Physical Layer (PHY) Specifications for Low Rate
11 Wireless Personal Area Networks (WPANs), IEEE
Standard 2006.
Pe ( 1 / 100)

10

9
5. Fitts correlation tracker simulation Control Lab. http://
8
webee.technion.ac.il/control/info/Projects/ Students/2008
/Shai Levi and NoamHelmer/Book/final_book.pdf. Date
7
accessed: 12/04/2016.
6
6. Kaur N, Kalsi A P S. Implementation of Polar Codes over
5
AWGN and Binary Symmetric Channel, Indian Journal of
4
Science and Technology. 2016 May, 9(19), pp .1–4.
3
7. Kadali K S, Rajaji L, Vijayaragavan, S P. An Efficient
2
Approach for the Removal of Bipolar Impulse Noise using
1
Median Filter, Indian Journal of Science and Technology.
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 X 2015 Jul, 8(13) pp .1–3.
SNRo ( dB) 8. Kingsbury N. Complex wavelets for shift invariant
analysis and filtering of signals, Journal of Applied and
Figure 8.  Graph showing Probability of error Vs (SNR)O Computational Harmonic Analysis. 2011, 10(3), pp.
for DSSS BPSK signal 234–253.
9. Geudtner D, Schwabisch M, Winter R. SAR-interferometry
with ERS-1 data, Proc. PIERS. 1994, pp. 11–15.
6. Conclusion 10. Candeias ALB, Dutra LV, Moreira JR, Mura JC.
Interferogram phase noise reduction using morphologi-
In this research work, DSSS system has been simulated cal and modified median filters, Proc. IGARSS. 1995, pp.
using coherent BPSK modulation/demodulation and 166–168.
matched filter. Here, we discussed and explored the use 11. Lee JS. Digital image enhancement and noise filtering by
of noise rejecting property of matched filter in DSSS sys- use of local statistics, IEEE Trans. Pattern Analysis Machine
tem for signal security. From the Tables 1 & 2, SNRs for Intell. 1980, 2(2), pp. 165–168.

6 Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology
 Iffat Rehman Ansari, Shahnawaz Uddin and Sameena Naaz

12. Yu Q, Yang X, Fu S, Liu X, Sun X. An adaptive contoured 16. Islam M S, Han H, Lee J I, Jung M G, Chong U. Small tar-
window filter for interferometric synthetic aperture radar. get detection and noise reduction in marine radar systems,
IEEE Geoscience and Remote Sensing Letters. 2007, 4(1), Proceedings of International Conference on Electronic
pp. 23–26. Engineering and Computer Science (EECS 2013). Science
13. Meng D, Sethu V, Ambikairajah E, Ge L. A novel technique Direct IERI Procedia. 2013, 4, pp. 168–173
for noise reduction in InSAR images, IEEE Geoscience and 17. Anjali M, Nelleri A. Phase-only Wavelet Matched
Remote Sensing Letters. 2007, 4(2), pp. 226–230. Filtering for Two- Dimensional Image Recognition and
14. Grispino AS, Petracca GO, Domínguez AE. Comparative Classification, Indian Journal of Science and Technology.
analysis of wavelet and EMD in the filtering of radar signal 2016 Sep, 9(36), pp .1–6.
affected by brown noise. IEEE Latin America Transactions. 18. Signal-Space Analysis: http://www.colorado.edu/engineer-
2013, 11(1), pp. 81–85. ing/ECE/ecen4242/ch05.ppt. Date Accessed: 14/05/2016.
15. Govoni MA, Hongbin LI. Range-Doppler resolution of the 19. Properties of a matched filter: https://www.coursehero.com/
linear-FM noise radar waveform, IEEE Transactions on file/p4cn9im/Properties-of-a-Matched-Filter-We-note-that-
Aerospace and Electronic Systems. 2013, 49(1), pp. 658-664. a-filter-which-is-matched-to-a/. Date Accessed: 20/5/2016.

Vol 9 (47) | December 2016 | www.indjst.org Indian Journal of Science and Technology 7