,QWHUQDWLRQDO:RUNVKRSRQ$QWHQQD7HFKQRORJ\L:$7

MIMO FM-CW radar using beat signal averaging

method
Nur Emileen Abd Rashid
Asiah Maryam Md Noor
Faculty of Electrical Engineering
Faculty of Electrical Engineering
Universiti Teknologi MARA
Universiti Teknologi MARA
Shah Alam, Selangor, Malaysia
Shah Alam, Selangor, Malaysia
emileen98@uitm.edu.my
asiahmaryamn@gmail.com
Idnin Pasya Raja Syamsul Azmir Raja Abdullah
Microwave Research Institute Faculty of Engineering
Universiti Teknologi MARA Universiti Putra Malaysia
2020 International Workshop on Antenna Technology (iWAT) 978-1-7281-2640-1/20/$31.00 ©2020 IEEE 10.1109/iWAT48004.2020.1570618451

Shah Alam, Selangor, Malaysia Selangor, Malaysia.

idnin@uitm.edu.my r_syamsul@upm.edu.my

Abstract— This paper proposes a MIMO FM-CW radar high resolution detecting ability despite the high clutter
system using beat signal averaging method for maritime interferences from the sea waves.
surveillance application. The system operates by transmitting
multiple waveform signals and receiving multiple waveform The implementation of multiple-input-multiple-output
signals, before averaging the beat signals at each receiver. A (MIMO) system in FM-CW radars are being studied as the
numerical simulation was conducted to validate the alternative to the conventional SISO system, as it may
performance of the proposed method in terms of ranging error provide better power allocation optimization, detection
probability vs SNR. It is found that the radar shows high probability against clutter, higher angular resolution, and
accuracy in range estimation even in low SNR values, increased accuracy when detecting multiple targets. These
compared to conventional FM-CW radar using single antenna advantages are made possible by the employment of
system. orthogonal waveforms through multiple antennas hence
provides waveform diversity, which makes it superior
Keywords—FMCW radar, MIMO, frequency diversity, radar compared to the existing SISO systems[3].
performance, maritime radar
This paper presents a method to implement MIMO in
I. INTRODUCTION FM-CW radar by averaging the beat signals at the receivers
and to study the performance of the system for maritime
Over the years, radar system for maritime surveillance application.
has attracted numerous research interests because of the
broad range of application for this technology, which
II. PROPOSED SCHEME: MIMO FM-CW RADAR BY BEAT
amongst them includes maritime law enforcement, search
SIGNAL AVERAGING
and rescue, environmental protection and resource
management [1]. A conventional maritime radar system
A. FM-CW radar and its MIMO configuration used in this
usually consists of on-shore located radars, with co-located
transmitting and receiving antenna (also denoted as “single- study
input-single-output”, SISO) configuration. Marine vessels In FM-CW radars, target range is estimated through the
are tracked by obtaining their distance and angle from the measurement of frequency shift between the transmitted
radar, by means of analysing the receiving radar signals signal and the received signal. The transmitted signal can be
reflected from them, while rotating the antenna beam denoted as [4] ;
direction. The target’s range from the radar often determined
by determining the time of arrival of reflected signal. This
can be easily done by using a pulse-based radar signal, or by (1)
modulating a continuous wave (CW) with a limited pulse
duration. A more robust radar signalling method is by using a where the signal is a sinusoidal wave in which its frequency
frequency modulated CW signal (FM-CW), which is linearly swept from f0 to f0+B in a chirp period T.
traditionally estimates the target range by producing
examining the beat frequency obtained from mixing the
receiving waveform with a reference signal. Another
advantages of the FM-CW radar system is its ability to not
only detect the range of a target, but also its velocity by
measuring the Doppler’s frequency of the moving target [2].
Radar systems in the conventional maritime monitoring
systems still require improvements in its physical limitations
which includes the signal propagation distance, the
resolution of detection, performance degradation due signal
interferences such as the sea clutter and spectrum occupancy
by other users. This problem becomes more severe when the
radar is detecting smaller targets such as small boats and Figure 1: Frequency-time expression of triangular
vessels. Hence it is required for a maritime radar to have swept FMCW signal

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

k,(((
Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 01:57:28 UTC from IEEE Xplore. Restrictions apply.
 ,QWHUQDWLRQDO:RUNVKRSRQ$QWHQQD7HFKQRORJ\L:$7

To use MIMO concept in this study, the frequency
diversity method was chosen in which the signals are swept
in different frequency bands. The transmitting signals for
this approach of MIMO FM-CW radar can be expressed as
per follows;
where is the propagation speed, T is the period for one
modulation cycle, B is the bandwidth of the signal and fb is
the beat frequency. Fig. 6 depicts the block diagram of the
proposed scheme.

(2)

where M is the total number of signals used in the radar.

Fig. 3. Range estimation of the SISO FM-CW system in different SNR

values

Fig. 2. Transmitted signals with frequency diversity in an FM-CW

radar by triangular sweep

The MIMO topology used in this study is MIMO 2×2

and MIMO 3×3 where each of the signal frequency bands
are defined as per below:
s
• 2×2: 0-10Fig., 11-21Fig. Fig. 4. Range estimation of the proposed MIMO FM-CW scheme (2×2) in
• 3×3: 0-10Fig., 11-21Fig., 22-32Fig. different SNR values

The signal sampling frequency used for the simulation

was 80 Fig.. The signals’ frequency is set to sweep in
triangular motion in a period of 10 ms where the bandwidth
of the signal, B was made constant at 10 Fig.. Band gaps of
1 Fig. between each frequency bands are applied as to avoid
the overlapping of the signals[5]. The transmission of the
signals is done in a monostatic, free-space radio
transmission. It was simulated that the radar’s target has an
radar cross section (RCS) of 20m2 and the reflection
coefficient of the target was simulated using Swerling
Model type 1, at 30 m distance from the radar in which the
distance is manipulated through the delay period of the
reflected signal. Additive White Gaussian Noise is included
Fig. 5. Range estimation of proposed MIMO FM-CW scheme (3×3) in
into the reflected signal at each receiver. At the receiver, the
different SNR values
signals are isolated by using a window-based filter design
with filter order = 58. Each filtered sub-band signals are
then mixed with respective sub-band reference signal to
produce the beat signals. The beat signals were averaged
Table 1: Average values of the estimated range deviation from the actual
before implementing peak detection through fast Fourier range
transform (FFT). The range of the target, detected by the
SNR SISO MIMO 2x2 MIMO 3x3
radar, R, was calculated using the following formula;
10d B 16.125m 8.500m 5.125m
5 dB 19.750m 19.187m 19.187m
(3)
0 dB 32.125m 19.375m 17.625m

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 01:57:28 UTC from IEEE Xplore. Restrictions apply.
 ,QWHUQDWLRQDO:RUNVKRSRQ$QWHQQD7HFKQRORJ\L:$7
Fig. 6. Simulation flow of proposed MIMO FM-CW radar scheme
Figs. 3, 4 and 5 show the numerical simulation result of
range estimation for target at 5 to 200 m by the proposed
radar. The results show that the MIMO radar produces
minimal deviation from the actual target range, in the case
of SNR = 10 dB, the range errors are deviated at an average
of 5.125m for 3×3 scheme and 8.5m for the 2×2 scheme
whereas in SISO the range errors are deviated from the
actual range at an average of 16.125m. Through this result,
it can be concluded that MIMO provides better estimation
result compared to SISO system as the increase in number
of receivers might help average out the noise and increase
the data reliability.
B. Ranging Performance study
The target range estimated by the radar was compared to
the actual distance of the target where the radar’s range
estimation error is defined as
(4)
Radar’s performance for both 2×2 and 3×3 MIMO
topologies were studied in terms of probability of range error
against SNR.
III. PROBABILITY OF RANGE ERROR AGAINST SNR
Fig. 7. Performance of proposed radar model in terms of
probability of range error against SNR
Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 01:57:28 UTC from IEEE Xplore. Restrictions apply.

Fig. 7 shows the radar’s performance in terms of
probability of ranging error against SNR for MIMO 2×2 and
MIMO 3×3, with the presence of an interval band (band
size=1Fig.). Both MIMO systems produces better results
than SISO in terms of robustness against noise. Observing at
0.2 error probability, the 2×2 MIMO improves by
approximately 8 dB and the 3×3 MIM) by 10 dB, compared
to the SISO case. This is attributable to the frequency
diversity brought by the MIMO configuration in detecting
Swerling-like targets.
IV. SUMMARY
In this paper, we proposed a method to implement
MIMO scheme in FM-CW radar by beat signal averaging
method. A numerical simulation was conducted to study the
proposed system, were the FM-CW radar signal was
transmitted in multiple waveforms and the target range was
estimated by averaging the beat signals from each receiver. It
is found that the radar was high in performance of range
estimation and high robustness against noise. The proposed
system will be further studied through experimental
validation and higher scheme of MIMO.
ACKNOWLEDGMENT
The authors thank the Collaborative Research in
Engineering, Science and Technology (CREST) for
providing the funds to support the present work under
CREST R&D Grant, project T05C1-67.
REFERENCES
[1] S. Maresca, P. Braca, J. Horstmann, and R. Grasso, “Maritime
surveillance using multiple high-frequency surface-wave radars,”
IEEE Trans. Geosci. Remote Sens., vol. 52, no. 8, pp. 5056–5071,
2014.
[2] M. I. Skolnik, Introduction to radar systems. McGraw-Hill, 1962.
[3] W. Q. Wang, H. C. So, and A. Farina, “An Overview on
Time/Frequency Modulated Array Processing,” IEEE J. Sel. Top.
Signal Process., vol. 11, no. 2, pp. 228–246, 2017.
[4] K. W. Gurgel, G. Antonischki, H. H. Essen, and T. Schlick, “Wellen
Radar (WERA): A new ground-wave HF radar for ocean remote
sensing,” Coast. Eng., vol. 37, no. 3–4, pp. 219–234, 1999.
[5] S. Zainuddin et al., “Performance of MIMO FMCW Radar in
Detecting Small Vessels,” 2018, pp. 2–5.