Received 23 August 2023, accepted 5 September 2023, date of publication 12 September 2023,

date of current version 20 September 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3314577

A Compact Textile Monopole Antenna

for Monitoring the Healing of Bone
Fractures Using Un-Supervised Machine
Learning Algorithm
SHAIK RIZWAN 1 , KANAPARTHI V. PHANI KUMAR 1, (Member, IEEE),
AND ABDULLAH J. ALAZEMI 2 , (Member, IEEE)
1 Diagnostics,Therapeutics and Assistive Devices Laboratory, Department of Electronics and Communication Engineering, Faculty of Engineering and
Technology, SRM Institute of Science and Technology, Kattankulathur, Tamil Nadu 603203, India
2 Department of Electrical Engineering, Faculty of Engineering and Petroleum, Kuwait University, Kuwait City 13060, Kuwait

Corresponding authors: Abdullah J. Alazemi ([email protected]) and Kanaparthi V. Phani Kumar ([email protected])
This work involved human subjects or animals in its research. The authors confirm that all human/animal subject research procedures and
protocols are exempt from review board approval.

ABSTRACT In this paper, a novel approach to monitoring the healing process of bone fractures
using a compact textile monopole antenna integrated into a wearable device. A planar wearable textile
ultra-wideband monopole (TUM) antenna is proposed as a sensing element. The antenna is fabricated on
the jeans fabric with a thickness of 1 mm, an overall size of 18 × 19 mm2 , and operates in an ultra-wideband
frequency range (3.22-10.9 GHz). The human arm tibia model is developed in a full-wave simulator in order
to estimate the radiation exposure caused by the antenna at near and far field distances. At both field distances,
the average specific absorption rate (SAR) values of the model are below the FCC limit. The proposed
antenna is experimentally tested on a donor bovine tibia. The proposed system utilizes Unsupervised machine
learning (ML) techniques to analyze the received signals from the antenna and provide real-time feedback
on the progression of bone fracture healing. The compact and flexible nature of the textile antenna allows
for comfortable and unobtrusive wear, making it suitable for long-term monitoring. The integration of
Unsupervised ML techniques enables automated analysis of the received signals, eliminating the need
for manual interpretation. This work aims to improve the monitoring and assessment of bone fracture
healing, leading to more effective treatment strategies and faster recovery times. The experimental results
and comparison study with previous literature show that the proposed TUM antenna has several advantages
and is suitable for diagnosing bone fractures.

INDEX TERMS Textile, monopole, ultra-wideband (UWB), bovine tibia, un-supervised machine learning.

I. INTRODUCTION issue, and insufficient or delayed treatment can result in

Bone fractures are common injuries that require careful a variety of issues, including bone weakening, unnatural
monitoring to ensure proper healing. Post-fracture recovery healing, and function loss. The tibia is the longest and
is a significant therapeutic burden worldwide that leads to most frequently fractured bone. Typically, tibial fractures
reduced mobility and impaired musculoskeletal function in a recover faster with supervised mobility [1], [2]. Early fracture
healthy body. It is considered a challenging musculoskeletal diagnosis is recommended because the unsupervised and
uncontrolled movement of the broken area may worsen the
The associate editor coordinating the review of this manuscript and injury. To detect tibial bone fractures, bone scans may require
approving it for publication was Tutku Karacolak . dye injection and repeated scanning [3]. Several integrated
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
VOLUME 11, 2023 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 101195
 S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

using the Un-supervised ML technique, PCA is reported

in [16], [17], and [18].
In [19], the authors have described the use of wide-band
textile antennas for microwave imaging and monitoring
neurodegenerative diseases. A miniaturized textile monopole
antenna is proposed in [20] to detect malignant tumors with-
out skin contact. The SAR is below FCC limitations, and the
operational frequency range is 8 to 12 GHz. In [21] and [22]
a purely textile antenna with an ultra-wideband for wearable
microwave imaging systems is proposed. A few textile anten-
nas for medical applications have been reported in [23], [24],
and [25]. Therefore, the textile monopole antennas are simple
to design and fabricate, small in size, low in cost, simple
FIGURE 1. The proposed TUM antenna design layout in full-wave to integrate with RF/microwave devices, and suitable for
simulator. detecting abnormalities in the human body.
In this work, we propose a compact textile monopole
antenna that can be integrated into a wearable device to
stress sensors were used to measure tibia bone damage [4]. monitor the healing process of bone fractures. The proposed
The emergence of wearable technology and wireless sensing antenna is made of widely available jeans cloth and conduc-
opens up new possibilities for continuous monitoring of bone tive copper and is based on a simple monopole structure.
fractures. The non-invasive microwave testing accelerates The proposed antenna has been tested experimentally on a
defect detection mechanisms in various fields [5], [6], [7]. donor bovine bone. Principal component analysis (PCA),
In recent years, microwave antennas for non-destructive an Unsupervised ML-based technique, is applied to the test
testing have attracted immense interest and can also identify results to detect bone fractures and monitor the healing status.
deep subterranean defects [8], [9]. Microwave systems are This manuscript is divided into four sections. Section I
non-ionizing and safer than other non-invasive testing meth- discusses tibia fractures, microwave antennas for medical
ods. Due to their low cost, low profile, and high performance, applications, and their features and challenges. The antenna
microwave systems have become an appropriate solution for design analysis, fabrication process, and performance anal-
medical diagnosis. The main integrated parts of microwave ysis are described in Section II. Section III explains the
systems are antennas. The antennas are used to detect changes methodology for detection and a comparison with prior
in the biological tissue’s permittivity. Due to the insulating works, while Section IV provides a conclusion
properties of cell membranes, permittivity is typically high
at lower frequencies and decreases at higher frequencies II. ANTENNA DESIGN DESCRIPTION
due to scattering effects [10]. Antennas with monopole A. DESIGN ANALYSIS
structures are another popular type that has received much 1) ANTENNA DESIGN
investigation for biomedical applications. This is due to the The Finite Integration Technique (FIT) based full-wave
fact that monopole structures may operate over a broad simulator is used to design and simulate the proposed TUM
frequency range with a compact size [11]. Several monopole antenna. The overall size of the structure is 18 × 19 × 1 mm3 ,
antennas have already been proposed for medical diagnostic which is equivalent to 0.186λl × 0.196λl × 0.011λl , where
applications in the literature. λl is the lower cut-off wavelength of 0.097m. The final layout
In [12], a portable microwave system to detect bone frac- of the proposed TUM antenna is shown in Fig. 1, and its
tures using a planar Electromagnetic Band Gap (EBG) sensor structural dimensions are SL = 19, SW = 18, CR = 1.9,
in contact with the patient’s skin. However, non-uniform R = ML = 3, CF = 0.75, E1 = 6.5, E2 = 2.4, FL = 6.55,
skin contact between successive measurement spots may FW = 3.4, GR = 2.9, GL = 6, and MW = 3.6 (All parameter
impact the measurement’s accuracy. In [13], the restoration values are in mm). The evolution of the design begins with
of bone profiles is examined using fabricated symmetrical the monopole structure depicted in Fig.1, which consists of
phantom geometries composed of the tibia, fibula, muscle, a circular radiation patch and a 50--microstrip feed line.
and fat. In [14], a loop-dual monopole antenna is used to Using the formulas given in [26] and [27], the equation for
detect heart failures. However, the resonator’s loop is used lower cut-off frequency (fl ) is formulated as follows,
to optimize the size of the antenna. Therefore, employing
7.2
a dual monopole structure increases the antenna profile, fl = (inGHz) (1)
which makes it less suitable for a wearable system. In [15], (l + a + d) × α
a monopole antenna with a modified or defective ground where l = 6 mm, is the height of the patch, a = 1.5 mm,
structure (DGS) is proposed. The intricate structure of the is the effective radius of the circular patch, d = 0.62 mm,
DGS creates difficulties in fabricating composite materials is the gap between the circular patch and the ground
and systems. The microwave antennas detect breast tumors plane, and the empirical value of substrate α = 1.23.

101196 VOLUME 11, 2023

S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

FIGURE 2. Evolution steps of proposed TUM antenna.

FIGURE 3. The |S11 |-parameters for the evolution steps of the proposed
TUM antenna.

Then, these parameters are fine-tuned in the full wave sim-

ulator to get a better |S11 |- curve. The operating bandwidth
(i.e., |S11 | < −10 dB) of the basic monopole antenna is from
5.15 to 6.23 GHz and 7.32 to 12 GHz. Therefore, the initially
designed primary monopole antenna cannot be characterized
as a wide-band antenna due to a notch in the middle of
two frequency bands and inadequate impedance matching.
Therefore, the structure of the patch and ground of the basic
monopole antenna is modified to achieve the desired ultra-
wideband [28], [29].

2) EVOLUTION OF THE PROPOSED TUM ANTENNA

The proposed TUM antenna evolution is depicted in Fig. 2.
The antenna is developed in five stages. In the first stage, FIGURE 4. The |S11 |−response variation with respect to (a) CR value
(b) ML (MW = 3.6 mm).
a circular-shaped patch monopole structure with microstrip
line feeding is designed (shown in Fig. 2(a)). However, at this
stage, the simple monopole antenna did not cover the UWB. semi-circular-shaped slots are introduced to the left and right
Therefore, the electrical length of the patch was extended sides of the integrated patch to achieve proper impedance
by convolving three elliptical-shaped resonators on the left matching. Finally, to achieve UWB, the ground plane
and right with an angle of 35, 45, and 55 degrees from structure is modified, and the feed is tapered, as shown in
the bottom of the circular patch, as shown in Fig. 2(b). Fig. Fig. 2(d) and 2(e), respectively. Therefore, the resulting
As a result, a wide band ranging from 3.54 to 12 GHz is bandwidth ranges from 3.12 to 10.86 GHz. Fig. 3 depicts the
obtained in the second stage. As illustrated in Fig. 2(c), the |S11 |− results for the antenna’s evolution steps.

VOLUME 11, 2023 101197

 S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

3) PARAMETRICAL STUDIES
In the second step of evolution, the |S11 | has a wider response
with −15 to −10 dB of the reflection coefficient. In order
to improve the response of the |S11 |− curve, two semi-
circular-shaped slots are introduced to the right and left sides
of the basic resonator. The radius of the semi-circular slot
is represented by the parameter CR , which is employed in
parametric optimization. This is accomplished by adjusting
the CR value in the full wave simulator within a specified
range. Fig. 4(a) depicts the |S11 |−response for CR value
variation. The simulated |S11 |−response obtained from the
second step of evolution is compared with the |S11 |−response
for the variation of the CR value. Finally, CR = 1.85 mm
is selected, and it gives an improvement in the reflection
coefficient.
However, certain |S11 |−responses are still near the
−10 dB line in step three evolution. Therefore, a U-slot
is introduced in the middle of the ground plane. Two
parameters, ML and MW of U-slot, are used for the parametric
optimization. By selecting MW = 3.6 mm and varying
the ML value from 1 to 2.5 mm, there is a shift in the
|S11 |−response. The modified |S11 |−plot with respect to
ML and MW is plotted in Fig. 4(b). It indicates that the
U-slot has a significant effect on S11 magnitude for the
band of interest. In the next step, to achieve the desired
S11 magnitude with an acceptable tolerance below the
−10 dB line, the ground plane is further modified with
the two semi-circular slots on the right and left sides.
At the final stage, the feed is tapered for the fine-tuning
of the |S11 |−response.The proposed TUM antenna achieves
FIGURE 5. (a) The fabricated model of proposed TUM antenna
effective impedance matching over the UWB frequency (b) Comparison of the measured frequency response with the full-wave
range, with resonant frequencies at 4.1, 7.5, and 9.6 GHz, simulated results.

respectively.
designed room that absorbs reflections and echoes, known
as an anechoic chamber. These measurements are aimed at
B. ANTENNA RESULT ANALYSIS capturing the far-field radiation patterns of the antenna, and
1) FREE-SPACE RESULTS they are compared against a standard horn antenna, which
The jeans cloth is used as the dielectric substrate of an serves as the reference. Fig. 7 illustrates the setup used for
antenna, and its properties are permittivity (ϵr ) is 1.7, tan δ measurements in the anechoic chamber, featuring the pro-
is 0.025, and a thickness of 1 mm. Following that, copper is posed TUM antenna under examination. The fields radiated
used for an antenna’s ground plane and radiation patch. The by the proposed TUM antenna in free space are visualized
fabricated model of the proposed TUM antenna and its |S11 |- in the azimuthal plane (H-plane) and the elevation plane (E-
response for the simulated and fabricated model is plotted in plane) in Fig. 7 for three different frequencies. The azimuthal
Fig. 5(a) and 5(b). The measured operating frequency range plane’s radiation pattern is donut-like (omnidirectional),
in free space is 3.22 to 10.9 GHz. In an anechoic chamber, the indicating that the antenna radiates energy in a broad circular
gain and radiation pattern are measured at different operating pattern around its axis. This circular pattern remains relatively
frequencies in free space. consistent across the three different frequency ranges. In the
Fig. 6 shows the complete competence of an antenna elevation plane, the radiation pattern of the proposed
analyzed using gain vs. the frequency (2 to 12 GHz). As the TUM antenna is a vertically polarized bi-directional struc-
frequency ranges between 3.22 to 10.9 GHz, the simulated ture. A quarter-wavelength proposed monopole design and
realized gain from 1.9 to 4.42 dBi is achieved, and the antenna element arrangement are significantly responsible
measured average gain is 3.83 dBi. The simulated radiation for achieving vertical polarization, intended for the proposed
efficiency of the antenna achieved ranges from 91.2 to wearable application. And also noticed that the radiation
95.86 % for an operating range of 3.12 to 10.86 GHz. patterns of the proposed TUM antenna for the upper band
The radiation properties of the proposed TUM antenna frequency are slightly different, and this may be due to the
are evaluated by conducting measurements in a specially change in surface current distributions at higher frequencies.

101198 VOLUME 11, 2023

S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

signals to establish patterns and correlations between the

signal characteristics and the stage of bone fracture healing.
The PCA simplifies high-dimensional data while preserving
similarities and summarizes features by minimizing the total
dimensions [32]. The trained algorithms can then analyze
real-time signals received from the antenna and provide
accurate feedback on the healing progress.
The data set is considered as the reflection coeffi-
cients for 1001 frequency points of a fractured bone with
(F-Curve) and without (N-Curve) blood clot. The 1001 reflec-
tion coefficients are the variables, and the N and F curves are
the two observations. Hence, the tumor can be predicted by
plotting the two coordinates representing the two points in
1001-dimensional space. If the observations (N and F curves)
FIGURE 6. The Realized gain and Radiation Efficiency of the proposed
have any correlation, the correlated points are observed in
TUM antenna. the 1001-dimensional space. In order to better understand the
PCA method, it is represented in terms of its linear algebra
C. SAR RESULTS fundamentals [33], and it is assumed that the data set is
The human safety of the proposed TUM antenna is assessed represented by an m × n matrix, P, where the m (1001)
by calculating the average SAR using a full wave sim- rows are the variables (reflection coefficients) and the n (2)
ulator in line with the IEEE Std. C95.1TM-2019 SAR columns are the observations.
criteria [30], [31]. The SAR value for the human arm bone  
SN 1 SF1
cubical model is calculated for both near-field (10 mm) and  SN 2 SF2 
far-field (18 mm) distances. The Average SAR value for input
P= . .
 
  (2)
power (Pi−SAR ) of 0 dBm at a near-field distance is 0.287,  . .


1.145, and 2.798 (W/kg) (10g), at 4.1, 7.5, and 9.6 GHz,
SN 1001 SF1001
respectively. When the antenna is placed at a far-field distance
with a Pi−SAR of 0 dBm, the SAR distributions are 0.085, The SN and SF are reflection coefficients belongs to N and
0.574, and 1.677 (W/kg) (10g), at 4.1, 7.5, and 9.6 GHz, F curves. The linear transformation and variance property
1
respectively. (σX2 = n−1 XX T ) is applied to the matrix P, and then the
The maximum allowable transmit powers and Average resultant matrix QR is,
SAR for a 10g tissue at a far field distance were calculated 1
in this study. The SAR distribution to Pi−SAR at a far-field QR = RRT (3)
n−1
distance is plotted in Fig. 8. This indicates that the antenna’s
R1 .RT1 . . . R1 .RT1001

transmission power of 10 dBm is safe. The safety evaluation  R2 .RT
1  1 . . . R2 .RT1001 
results suggest that the proposed TUM antenna is suitable
. . . . .

QR =   (4)
for medical diagnosis. This study also includes a detailed n−1 . . . . .
 

comparison between the proposed TUM antenna and other
reported antennas. To ensure comparability, the SAR for R1001 .RT1 . . . R1001 .R1001
T

the proposed antenna is calculated by keeping a minimum where R, is the linear transformation of matrix P. Eigen values
distance of 5mm and providing an input power of 30 dBm. and Eigen vectors (E) are determined using the diagonal
For this case, the obtained SAR value is 1.13 (W/kg) for 1 gm matrix entries of QR , and X is a symmetric matrix, which is
of tissue. Therefore, even after keeping a minimum distance used find the correlation between the two curves. Therefore,
and giving maximum input power, the obtained SAR value the covariance matrix Cv is,
using the proposed TUM antenna is better when compared to  
C1N C1F
other works.  C2N C2F 
1
EXE =  .
T
. 
 
III. EXPERIMENTAL INVESTIGATION OF HEALING OF
Cv =  (5)
n−1  . . 

BONE FRACTURES
C1001N C1001F
A. METHODOLOGY
The compact textile monopole antenna is designed and where C1N and C1F are the first principal components
fabricated using textile materials, allowing it to be seamlessly (highest feature) of the N and F curves. Using equation (5),
integrated into clothing or bandages. The antenna operates the variances can be automatically extracted from the data
in the wireless frequency range suitable for transmitting and regarding the relative significance of each principal compo-
receiving signals from the fracture site. An Unsupervised ML nent. The highest variance is the first principal component,
technique, such as PCA, is trained on a dataset of received followed by the second highest principal component, and so

VOLUME 11, 2023 101199

 S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

FIGURE 7. The Radiation pattern of the antenna for three different frequencies.

Experiments are done on the different areas of the bone to

find out where the bone is broken and how well it is healing.
A more realistic evaluation is conducted using a bovine tibia
with a diameter of 3.8 cm and a length of 34 cm. In addition,
the bone is covered with muscle, fat, and skin that are 1.6 cm,
0.3 cm, and 0.8 cm thick. The experimental setup for bone
crack detection and its healing monitoring is depicted in
Fig. 9. The proposed TUM antenna is placed on the tibia with
muscle, fat, and skin over a far-field distance to account for
variations in the antenna’s radiating fields.
The steps involved in this experimental process are as
follows:
• Step 1: To begin the analysis, the proposed TUM antenna
is placed in front of the bovine tibia without fracture
over a far-field distance. The magnitude of S-parameters
is noted with 1001 frequency points in the operating
band.
• Step 2: A 3 mm crack is introduced in the tibia to
replicate a crack with a 35 mm blood clot, and the data
points are recorded in the same manner as in Step 1.
• Step 3: To monitor the fracture healing process, a blood
strip of dimension Lc × 10 × 5 mm3 is placed above
the fractured bone to mimic the healing model. The
blood strip length (Lc ) is varied to implicate the healing
status of the fracture. The parameter Lc is changed
from 35 mm to the lengths of 25, 15, and normal tibia
without fracture, and the data points for those lengths
FIGURE 8. The SAR distribution with respect to Pi −SAR at both near and are recorded.
far-field distances. Figures 10(a) and 10(b) show the measured S-parameters
of normal bone and fractured bone with varying lengths of
blood clots. From the reflection parameter plots, it is highly
on. Therefore, the above-described PCA approach is coded difficult to distinguish the bone with and without a crack and
in MATLAB and analyzed with the measured reflection its healing status. The data collected from the above steps are
coefficients. further analyzed using an Unsupervised ML technique, PCA.
The PCA simplifies high-dimensional data while preserving
B. EXPERIMENTAL ANALYSIS similarities and summarizes features by minimizing the total
A bovine tibia is used in the experiment since its dielectric dimensions [33]. PCA maps data onto principal components
properties are nearly equal to the human tibia [34], [35]. (PCs) in a geometric way to find the best data summary with

101200 VOLUME 11, 2023

S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

FIGURE 9. The Experimental setup for bone crack detection and its
healing monitoring.

a limited number of PCs. The accuracy of the approximation

can be evaluated based on the amount of variation associated
with the PCs. First, we apply PCA to the data points obtained
from the proposed TUM antenna with and without fractured
tibias, as illustrated in Fig. 10(a), and the PCA results are
plotted in Fig. 11(a). Where AN and A35−F are PCs derived
from reflection parameters of normal and fractured areas of
bone, respectively. These results demonstrate the distinction
between broken and normal bones.
Similarly, the same PCA is then applied to the reflection
parameters depicted in Fig. 10(b) to detect the blood clots on
the bone fractures of various lengths. The PC values B35−F
and B25−F are obtained by applying PCA to the reflection
parameters Lc = 35 mm and Lc = 25 mm and are plotted in
Fig. 11(b). From this first iteration, we can easily distinguish
blood strips by correlating the B35−F with the B25−F .
In the second step of iterations, the PC values, C25−F ,
and C15−F are generated by applying PCA to the reflection
parameters obtained for Lc = 25 mm, and Lc = 15 mm,
respectively, and are plotted in Fig. 11(c). In the next step FIGURE 10. (a) The |S11 |− parameters for normal and fractured areas and
(b) the blood clots’ lengths (35 mm, 25 mm, 15 mm) of the bovine tibia.
of iterations, the PC values, D15−F , and DN are generated
by applying PCA to the reflection parameters obtained for
Lc = 15 mm, and Normal tibia without fracture, respectively, The highlights of the proposed work are as follows,
and are plotted in Fig. 11(d). The PC values for each case • A non-invasive and non-ionized approach is developed
are different. Therefore, the blood clot, which forms during to detect and monitor bone fractures and their healing
fracture healing, can be easily distinguished by observing the process using a small-size planar monopole antenna.
PCA plots. • The proposed antenna prototype is tested on the bovine
Unsupervised ML techniques, PCA is used to detect and bone (with skin, fat, and muscle) and the effect of
monitor bone fractures based on the reflection parameters of different length blood clots with cracks is investigated.
the proposed TUM antenna. Therefore, the proposed TUM • The Unsupervised ML method is applied to reflection
antenna could be used effectively for bone fracture diagnosis. coefficient parameters to identify the presence of
Table 1 demonstrates that the proposed TUM antenna fracture and its healing.
has many advantages, including small or comparable size, • As compared to the previously reported works, the
adequate bandwidth, acceptable gain, and the lowest SAR proposed diagnostic system has several advantages
over traditional wearable antennas with similar dielectric such as a non-invasive mechanism, precise, portability,
constant [36], [37], [38], [39], [40] and other UWB antennas a skilled technician is not required, and a planar
realized using FR4 substrates [41], [42], [43]. Table 2 antenna (easy to integrate with RF/IoT devices) for crack
shows the comparison of the proposed antenna with other detection and its healing monitoring which is hardly
state-of-the-art literature. The proposed TUM antenna is reported.
compact, lightweight, inexpensive, biodegradable, wearable, The proposed system has been tested using experimental
conformal, and easy to fabricate and integrate with RF/IoT fracture healing scenarios and preliminary bovine bone trials.
wearable devices. The results demonstrate the feasibility and effectiveness of

VOLUME 11, 2023 101201

 S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

TABLE 1. Comparison of several advantages of the proposed TUM antenna over recent designs.

TABLE 2. Comparison of the proposed TUM antenna with other state-of-the-art literature for biomedical applications.

on the received signals. The system provides real-time

feedback to healthcare professionals, allowing them to track
the progress of fracture healing and adjust treatment plans
accordingly.

IV. CONCLUSION
This research work presents a compact textile monopole
antenna integrated into a wearable device for monitoring
the healing process of bone fractures. The proposed TUM
antenna has an average measured gain of 3.83 dBi and a much
higher operating bandwidth with a compact size. The antenna
operates in close enough proximity to the human body with
a near-field and far-field distance, and the average SAR is
below the FCC limit for all operating frequencies with an
input power of 0 to 10 dBm.
The integration of Unsupervised ML techniques enables
FIGURE 11. The PCA results for (a) Normal Tibia (NT) and Fractured automated analysis of received signals, providing real-time
Tibia (FT) with Lc = 35mm, (b) FT with Lc = 35mm and FT with
Lc = 25mm, (c) FT with Lc = 25mm and FT with Lc = 15mm, (d) FT with
feedback on fracture healing progress. The proposed sys-
Lc = 15mm and NT without fracture and blood clot. tem has the potential to revolutionize fracture monitoring
by offering a non-invasive, continuous, and cost-effective
using the compact textile monopole antenna in monitoring solution. This technology can significantly improve treatment
bone fracture healing. The Unsupervised ML technique strategies, reduce recovery times, and enhance patient care in
achieves high accuracy in predicting the healing stage based the field of orthopedics. Future work will focus on further

101202 VOLUME 11, 2023

S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

validation and optimization of the system for widespread [20] A. Mersani, L. Osman, and J.-M. Ribero, ‘‘Flexible UWB AMC antenna
clinical adoption. for early stage skin cancer identification,’’ Prog. Electromagn. Res. M,
vol. 80, pp. 71–81, 2019.
[21] X. Lin, Y. Chen, Z. Gong, B.-C. Seet, L. Huang, and Y. Lu, ‘‘Ultrawideband
REFERENCES textile antenna for wearable microwave medical imaging applications,’’
[1] J. Kenwright, J. B. Richardson, J. L. Cunningham, S. H. White, IEEE Trans. Antennas Propag., vol. 68, no. 6, pp. 4238–4249, Jun. 2020.
A. E. Goodship, M. A. Adams, P. A. Magnussen, and J. H. Newman, [22] H. Yalduz, T. E. Tabaru, V. T. Kilic, and M. Turkmen, ‘‘Design and analysis
‘‘Axial movement and tibial fractures. A controlled randomized trial of of low profile and low SAR full-textile UWB wearable antenna with
treatment,’’ Bone Joint J., vol. 73, no. 4, pp. 654–659, 1991. metamaterial for WBAN applications,’’ AEU-Int. J. Electron. Commun.,
[2] T. Miclau, C. Lu, Z. Thompson, P. Choi, C. Puttlitz, R. Marcucio, vol. 126, Nov. 2020, Art. no. 153465.
and J. A. Helms, ‘‘Effects of delayed stabilization on fracture healing,’’ [23] P. J. Soh, B. Van den Bergh, H. Xu, H. Aliakbarian, S. Farsi, P. Samal,
J. Orthopaedic Res., vol. 25, no. 12, pp. 1552–1558, Dec. 2007. G. A. Vandenbosch, D. M. P. Schreurs, and B. K. Nauwelaers, ‘‘A smart
[3] C. Milgrom, R. Chisin, M. Giladi, M. Stein, H. Kashtan, J. Margulies, and wearable textile array system for biomedical telemetry applications,’’ IEEE
H. Atlan, ‘‘Negative bone scans in impending tibial stress fractures,’’ Amer. Trans. Microw. Theory Techn., vol. 61, no. 5, pp. 2253–2261, May 2013.
J. Sports Med., vol. 12, no. 6, pp. 488–491, Nov. 1984. [24] M. M. H. Mahfuz, M. R. Islam, C. Park, E. A. A. Elsheikh, F. M. Suliman,
[4] G. Y. Yang, G. Johnson, W. C. Tang, and J. H. Keyak, ‘‘Parylene-based M. H. Habaebi, N. A. Malek, and N. Sakib, ‘‘Wearable textile patch
strain sensors for bone,’’ IEEE Sensors J., vol. 7, no. 12, pp. 1693–1697, antenna: Challenges and future directions,’’ IEEE Access, vol. 10,
Dec. 2007. pp. 38406–38427, 2022.
[5] N. Qaddoumi, A. H. El-hag, M. Al Hosani, I. A. Mansouri, and [25] I. Martinez, C. Mao, D. Vital, H. Shahariar, D. H. Werner, J. S. Jur,
H. A. Ghufli, ‘‘Detecting defects in outdoor non-ceramic insulators using and S. Bhardwaj, ‘‘Compact, low-profile and robust textile antennas with
near-field microwave non-destructive testing,’’ IEEE Trans. Dielectr. improved bandwidth for easy garment integration,’’ IEEE Access, vol. 8,
Electr. Insul., vol. 17, no. 2, pp. 402–407, Apr. 2010. pp. 77490–77500, 2020.
[6] H. Zhang, B. Gao, G. Y. Tian, W. L. Woo, and A. Simm, ‘‘Spatial-frequency [26] N. P. Agrawall, G. Kumar, and K. P. Ray, ‘‘Wide-band planar monopole
spectrum characteristics analysis with different lift-offs for microwave antennas,’’ IEEE Trans. Antennas Propag., vol. 46, no. 2, pp. 294–295,
nondestructive testing and evaluation using Itakura–Saito nonnegative Feb. 1998.
matrix factorization,’’ IEEE Sensors J., vol. 14, no. 6, pp. 1822–1830, [27] K. P. Ray and Y. Ranga, ‘‘Ultra-wideband printed elliptical monopole
Jun. 2014. antennas,’’ IEEE Trans. Antennas Propag., vol. 55, no. 4, pp. 1189–1192,
[7] A. J. Jundi and N. N. Qaddoumi, ‘‘Near-field microwave model of mul- Apr. 2007.
tilayered structures illuminated by open-ended rectangular waveguides,’’ [28] S. Doddipalli, A. Kothari, and P. Peshwe, ‘‘A low profile ultra wide
IET Microw., Antennas Propag., vol. 6, no. 1, pp. 100–107, 2012. band monopole antenna for wearable applications,’’ Int. J. Ant. Propag.,
[8] V. S. Ramalingam, M. Kanagasabai, and E. F. Sundarsingh, ‘‘Detection vol. 2017, Aug. 2017, Art. no. 7362431.
of voids in fiber reinforced plastics using magnetoinductive coupled [29] S. Rizwan, K. V. Phani Kumar, and S. K. Palaniswamy, ‘‘On the
microstrip sensor,’’ IEEE Sensors J., vol. 15, no. 8, pp. 4182–4183, experimental investigation of bone fracture recovery process using an ultra-
Aug. 2015. wideband planar monopole antenna,’’ Int. J. Antennas Propag., vol. 2023,
[9] M. H. Zarifi, S. Deif, M. Abdolrazzaghi, B. Chen, D. Ramsawak, pp. 1–13, Jul. 2023, Art. no. 8825446.
M. Amyotte, N. Vahabisani, Z. Hashisho, W. Chen, and M. Daneshmand, [30] A. Ahlbom, U. Bergqvist, J. H. Bernhardt, J. P. Cesarini, M. Grandolfo,
‘‘A microwave ring resonator sensor for early detection of breaches M. Hietanen, A. F. Mckinlay, M. H. Repacholi, D. H. Sliney, J. A. Stolwijk,
in pipeline coatings,’’ IEEE Trans. Ind. Electron., vol. 65, no. 2, and M. L. Swicord, ‘‘Guidelines for limiting exposure to time-varying
pp. 1626–1635, Feb. 2018. electric, magnetic, and electromagnetic fields (up to 300 GHz),’’ Health
[10] T. M. Grzegorczyk, P. M. Meaney, P. A. Kaufman, Phys., vol. 74, no. 4, pp. 494–522, 1998.
R. M. di Florio-Alexander, and K. D. Paulsen, ‘‘Fast 3-D tomographic [31] IEEE Standard for Safety Levels With Respect to Human Expo-
microwave imaging for breast cancer detection,’’ IEEE Trans. Med. Imag., sure to Electric, Magnetic, and Electromagnetic Fields, 0 Hz to
vol. 31, no. 8, pp. 1584–1592, Aug. 2012. 300 GHz, IEEE Standard C95.1-2019, 2019. [Online]. Available:
[11] D. M. Pozar, ‘‘Microstrip antennas,’’ Proc. IEEE, vol. 80, no. 1, pp. 79–91, https://standards.ieee.org/standard/C95 1-2019.html
Jan. 1992. [32] J. Lever, M. Krzywinski, and N. Altman, ‘‘Principal component analysis,’’
[12] V. S. Ramalingam, M. Kanagasabai, and E. F. Sundarsingh, ‘‘A compact Nature Methods, vol. 14, no. 7, pp. 641–642, 2017.
microwave device for fracture diagnosis of the human tibia,’’ IEEE Trans. [33] H. Abdi and L. J. Williams, ‘‘Principal component analysis,’’ Wiley
Compon., Packag., Manuf. Technol., vol. 9, no. 4, pp. 661–668, Apr. 2019. Interdiscipl. Rev., Comput. Statist., vol. 2, pp. 433–459, 2010.
[13] G. Ruvio, A. Cuccaro, R. Solimene, A. Brancaccio, B. Basile, and [34] B. Amin, M. A. Elahi, A. Shahzad, E. Porter, B. McDermott, and
M. J. Ammann, ‘‘Microwave bone imaging: A preliminary scanning M. O’Halloran, ‘‘Dielectric properties of bones for the monitoring of
system for proof-of-concept,’’ Healthcare Technol. Lett., vol. 3, no. 3, osteoporosis,’’ Med. Biol. Eng. Comput., vol. 57, no. 1, pp. 1–13, Jan. 2019.
pp. 218–221, Sep. 2016. [35] B. Amin, A. Shahzad, L. Farina, E. Parle, L. McNamara, M. O’Halloran,
[14] S. A. Rezaeieh, A. Abbosh, and Y. Wang, ‘‘Wideband unidirectional and M. A. Elahi, ‘‘Dielectric characterization of diseased human trabecular
antenna of folded structure in microwave system for early detection of bones at microwave frequency,’’ Med. Eng. Phys., vol. 78, pp. 21–28,
congestive heart failure,’’ IEEE Trans. Antennas Propag., vol. 62, no. 10, Apr. 2020.
pp. 5375–5381, Oct. 2014. [36] M. Mustaqim, B. A. Khawaja, H. T. Chattha, K. Shafique, M. J. Zafar,
[15] N. Ojaroudi, M. Ojaroudi, and N. Ghadimi, ‘‘UWB omnidirectional square and M. Jamil, ‘‘Ultra-wideband antenna for wearable Internet of Things
monopole antenna for use in circular cylindrical microwave imaging devices and wireless body area network applications,’’ Int. J. Numer.
systems,’’ IEEE Antennas Wireless Propag. Lett., vol. 11, pp. 1350–1353, Model., Electron. Netw., Devices Fields, vol. 32, no. 6, p. 2590, Nov. 2019.
2012. [37] S. Bhavani and T. Shanmuganantham, ‘‘Analysis of tumor in the breast
[16] M. A. Aldhaeebi, T. S. Almoneef, A. Ali, Z. Ren, and O. M. Ramahi, ‘‘Near using UWB textile antenna,’’ Microw. Opt. Technol. Lett., vol. 64, no. 10,
field breast tumor detection using ultra-narrow band probe with machine pp. 1765–1772, Oct. 2022.
learning techniques,’’ Sci. Rep., vol. 8, no. 1, pp. 1–16, Aug. 2018. [38] M. Kanagasabai, P. Sambandam, M. G. N. Alsath, S. Palaniswamy,
[17] M. K. Sharma, M. Kumar, J. P. Saini, D. Gangwar, B. K. Kanaujia, A. Ravichandran, and C. Girinathan, ‘‘Miniaturized circularly polarized
S. P. Singh, and A. L. Ekuakille, ‘‘Experimental investigation of the UWB antenna for body centric communication,’’ IEEE Trans. Antennas
breast phantom for tumor detection using ultra-wide band-MIMO antenna Propag., vol. 70, no. 1, pp. 189–196, Jan. 2022.
sensor (UMAS) probe,’’ IEEE Sensors J., vol. 20, no. 12, pp. 6745–6752, [39] A. Yadav, V. K. Singh, A. K. Bhoi, G. Marques, B. Garcia-Zapirain, and
Jun. 2020. I. de la Torre Díez, ‘‘Wireless body area networks: UWB wearable textile
[18] P. K. Rao and R. Mishra, ‘‘Resonator based antenna sensor for breast cancer antenna for telemedicine and mobile health systems,’’ Micromachines,
detection,’’ Prog. Electromagn. Res. M, vol. 101, pp. 149–159, 2021. vol. 11, no. 6, p. 558, May 2020.
[19] I. Saied and T. Arslan, ‘‘Wideband textile antenna for monitoring [40] A. Smida, A. Iqbal, A. J. Alazemi, M. I. Waly, R. Ghayoula, and S. Kim,
neurodegenerative diseases,’’ in Proc. IEEE 29th Annu. Int. Symp. Pers., ‘‘Wideband wearable antenna for biomedical telemetry applications,’’
Indoor Mobile Radio Commun. (PIMRC), Sep. 2018, pp. 356–360. IEEE Access, vol. 8, pp. 15687–15694, 2020.

VOLUME 11, 2023 101203

 S. Rizwan et al.: Compact Textile Monopole Antenna for Monitoring the Healing of Bone Fractures

[41] V. Sharbati, P. Rezaei, and M. M. Fakharian, ‘‘Compact planar UWB KANAPARTHI V. PHANI KUMAR (Member,
antenna with enhanced bandwidth and switchable band-notch function for IEEE) received the B.Tech. degree in electronics
WLAN and DSRC,’’ IETE J. Res., vol. 63, no. 6, pp. 805–812, Nov. 2017. and communication engineering from Acharya
[42] M. M. Fakharian and P. Rezaei, ‘‘Very compact palmate leaf-shaped CPW- Nagarjuna University, Guntur, Andhra Pradesh,
FED monopole antenna for UWB applications,’’ Microw. Opt. Technol. India, in 2010, and the M.Des. degree in
Lett., vol. 56, no. 7, pp. 1612–1616, Jul. 2014. electronic systems and the Ph.D. degree in
[43] M. B. Kakhki and P. Rezaei, ‘‘Reconfigurable microstrip slot antenna with electronics engineering from the Indian Institute
DGS for UWB applications,’’ Int. J. Microw. Wireless Technol., vol. 9, of Information Technology, Design, and Manufac-
no. 7, pp. 1517–1522, Sep. 2017.
turing, Kancheepuram, India, in 2013 and 2017,
[44] H. T. Sediq, J. Nourinia, C. Ghobadi, and B. Mohammadi, ‘‘A novel
respectively. He is currently a Research Assistant
shaped ultrawideband fractal antenna for medical purposes,’’ Biomed.
Signal Process. Control, vol. 80, Feb. 2023, Art. no. 104363. Professor with the SRM Institute of Science and Technology, Kattankulathur,
[45] A. S. M. Alqadami, N. Nguyen-Trong, A. E. Stancombe, K. Bialkowski, Tamil Nadu, India. His research interests include the design of microwave
and A. Abbosh, ‘‘Compact flexible wideband antenna for on-body passive circuits, RFIC, and rehabilitative and assistive technology.
electromagnetic medical diagnostic systems,’’ IEEE Trans. Antennas
Propag., vol. 68, no. 12, pp. 8180–8185, Dec. 2020.
[46] I. M. Danjuma, M. O. Akinsolu, C. H. See, R. A. Abd-Alhameed, and
B. Liu, ‘‘Design and optimization of a slotted monopole antenna for ultra-
wideband body-centric imaging applications,’’ IEEE J. Electromagn., RF
Microw. Med. Biol., vol. 4, no. 2, pp. 140–147, Jun. 2020.
[47] A. B. Dey, S. S. Pattanayak, D. Mitra, and W. Arif, ‘‘Investigation
and design of enhanced decoupled UWB MIMO antenna for wearable
applications,’’ Microw. Opt. Technol. Lett., vol. 63, no. 3, pp. 845–861,
Mar. 2021.
[48] X. Lin, Y. Chen, B.-C. Seet, F. Xia, Q. Zhang, and J. Hu, ‘‘Compact textile
wideband antenna for wearable microwave stroke imaging,’’ in Proc. 13th
Eur. Conf. Antennas Propag. (EuCAP), Mar. 2019, pp. 1–5.

SHAIK RIZWAN was born in Andhra Pradesh, ABDULLAH J. ALAZEMI (Member, IEEE)
India, in June 1991. He received the Bachelor of received the B.S. degree in electrical engineering
Technology (B.Tech.) degree in electronics and from Kuwait University, in 2010, and the M.S.
communication engineering and the Master of and Ph.D. degrees in electrical and computer
Technology (M.Tech.) degree in electronic sys- engineering from the University of California at
tems engineering from Jawaharlal Nehru Techno- San Diego, La Jolla, CA, USA, in 2013 and 2015,
logical University, Anantapur, India, in 2013 and respectively. His works focus on tunable antennas
2015, respectively. He is currently pursuing the and filters with RF-MEMS and MM-wave to THz
Doctor of Philosophy (Ph.D.) degree in RF and Quasi-optical systems.
microwave devices with the SRM Institute of He joined the Department of Electrical Engi-
Science and Technology, Kattankulathur, Tamil Nadu, India. He joined the neering, Kuwait University, in 2016, where he is currently an Associate
Department of Electronics and Communication, PBR Visvodaya Institute Professor. His works focus on 5G reconfigurable antennas for biomedical
of Technology and Science, as an Assistant Professor. In 2020, he joined applications, RF switches and antennas based on gap waveguide technology,
as a full-time Research Scholar to continue his study for the Doctor multiband power dividers, and couplers for advanced communication
of Philosophy (Ph.D.) degree in RF and microwave devices with the systems. He is currently the Director of the Microwave Engineering
SRM Institute of Science and Technology. His research interests include Laboratory, Kuwait University. In 2021, he received the Kuwait Award
printed antennas, flexible antennas, wearable antennas, microwave imaging, for excellence and creativity in science and technology provided by His
embedded systems, the Internet of Things (IoT), machine learning, and Highness Sheikh Nawaf Al Ahmad Al Jaber Al Sabah, Amir of Kuwait.
portable systems for biomedical applications.

101204 VOLUME 11, 2023