Overlapping-based Radio Signal Processing for SDF Location Method

Full text

Overlapping-based Radio Signal Processing
for SDF Location Method
Rafał Szczepanik
Institute of Communications Systems,
Faculty of Electronics,
Military University of Technology
Warsaw, Poland
rafal.szczepanik@wat.edu.pl
ORCID: 0000-0003-1225-297X
Jan M. Kelner
Institute of Communications Systems,
Faculty of Electronics,
Military University of Technology
Warsaw, Poland
jan.kelner@wat.edu.pl
ORCID: 0000-0002-3902-0784
Cezary Ziółkowski
Institute of Communications Systems,
Faculty of Electronics,
Military University of Technology
Warsaw, Poland
cezary.ziolkowski@wat.edu.pl
ORCID: 0000-0003-0750-0705
Abstract—The localization of radio emitters is commonly
used in the positioning of nodes, e.g., cellular network, radio
navigation, search and rescue operations, as well as in military
applications such as blue force tracking (i.e., positioning of own
troops), communication (COMINT) or electronic intelligence
(ELINT), and electronic warfare. Depending on the location
technique used, different signal processing methods are used. In
this paper, we present an efficient and adaptive signal
processing method for the signal Doppler frequency (SDF)
location method. This method is based on the variability of the
Doppler curve, the shape of which is distinctive for the mutual
position of the transmitter in relation to the receiver trajectory
or vice versa. The proposed approach is based on an overlap
technique that enables obtaining the desired number of Doppler
frequency shifts for the SDF. The effectiveness of this solution is
presented on the example of signals recorded in a real
environment. The proposed method could be used in signal
processing in real-time SDF applications as well as in simulation
studies.
Keywords—wireless communications, localization, signal
Doppler frequency (SDF) method, Doppler effect, overlapping
technique
I. INTRODUCTION
Modern wireless communication systems have primarily
contributed to the dynamic development of the existing and
the development of new location techniques [1,2]. Earlier,
these techniques were used mainly for the purposes of
radiolocation and radio reconnaissance in the armed forces
and in navigation systems used in air and sea transport. The
beginnings of mobile telephony forced the introduction of
location services (LCSs) for positioning subscribers for
network management needs. Over time, LCSs used by
operators evolved towards location-based services (LBSs)
provided by operators and service providers [3]. At LBSs, the
localization of radio facilities is only a supporting element for
the proper implementation of services and applications.
Navigation and sports applications, weather forecasts,
warning and alerting people in danger, local news and
information (i.e., location-based media), mobile games (i.e.,
location-based games), or other intelligent services (i.e.,
location intelligence, awareness) are examples of the use of
the location procedure in complex telecommunications
services [4,5].
Most of the location techniques rely on the emitter position
estimation based on the received signal parameters. One of the
most frequently implemented techniques is based on
measuring the directions of electromagnetic wave arrival, i.e.,
it comes down to determining the bearing to the radiating
source for several receiver (Rx) positions. Determining the
position in outdoor open or free spaces is relatively simple and
accurate. However, locating a Tx in an urban environment is
a much more complex task. This is due to the specificity of
propagation phenomena occurring in radio channels,
including multipath propagation, Doppler effect, dispersion of
the signal in the time, frequency, and reception angle domains.
To counteract, compensate, or mitigate these phenomena
effects, various techniques are used, such as coding, channel
equalizers, MIMO techniques, beamforming, and rake Rx.
The system solution is based on the use of a wireless sensor
network (WSN), the elements of which are responsible for the
individual location of the emission source. Then, a joint
decision about the positioned object is developed [2,6].
The paper aim is to present the method of adaptive signal
processing in the signal Doppler frequency (SDF) location
method [7]. The presented method applies to stationary or
slowly moving emission sources in relation to a mobile
receiver. In this technique, we used the distinctive features of
the Doppler curve, which are related to the position of the Rx
motion trajectory in relation to the object location. They are
the basis for determining the emitter coordinates. The paper
presents an effectiveness assessment of the overlapping
method [8] of measurement data in relation to direct signal
processing. The comparative analysis was carried out based
on the data obtained in a measurement company carried out in
a real environment.
The layout of the paper remainder is following. Section II
shows the short description of the SDF. Next, the overlapping
procedure used for the signal processing in this location
technique is presented in Section III. In Section IV, we will
show the effectiveness evaluation of the developed procedure
based on signals recorded during the empirical studies. The
conclusion of the paper is contained in the final section.
II. SIGNAL DOPPLER FREQUENCY METHOD
The solution of the wave equation, which considers the
mobility of the emission source, is the basis of the SDF. This
solution describes the relationship between the Doppler
frequency shift (DFS) and the position coordinates of the Tx
and Rx [9]:
( ) ( ) ( )
0
02
,, ,
1
D
fk x vt
f t f t f k rt
k

−
= − = +

−

xx x
(1)
where
0
f
and
( )
tf ,x
are carrier frequency and temporary
frequency of the received signal, respectively,
This work was co-financed by the Military University of Technology
under Research Project no. UGB/22-854/2021/WAT on "Applications of
selected computer science, communication, and reconnaissance techniques
in civil and military areas" and the Polish Ministry of Defense under
Research Grant no. GBMON/13-996/2018/WAT on "Basic research in
sensor technology field using innovative data processing methods".
978-0-7381-1340-1/21/$31.00 ©2021 IEEE 268 SPSympo 2021

,k v c=
c
and
v
are the light and Rx speeds, respectively,
( )
,,x y z=x
represents the emission source coordinates, and
( ) ( )
( )( )
22 2 2
, 1 .r t x vt k y z= − + − +x
The emitter position estimation is based on (1). This
procedure consists of measuring the DFS,
,
D
f
in successive
receiver positions. The assumption that the emission source is
located on the earth surface (
0z=
) allows us to determine its
coordinates
( )
,xy
based on the measurement of
D
f
in two
successive time moments, i.e.,
1
t
and
2,t
( ) ( )
( ) ( )
( ) ( ) ( )
( ) ( )
1 1 2 2
12
2
1 2 1 2 2
12
t A t t A t
xv
A t A t
t t A t A t
y v z
A t A t
−
=
−



−
=  −

−




(2)
where
( ) ( ) ( )
2
1,A t F t F t=−
( ) ( )
maxDD
F t f t f
and
max 0D
f f k=
are the normalized and maximum DFSs,
respectively.
In the case of a complex shape of the object's motion
trajectory, we separate rectilinear sections. The position of the
emitter is determined in the local coordinate system at
individual sections, and then the obtained result is transformed
into the global system. A characteristic feature of the SDF is
the possibility of the simultaneous location of many emitters
[10], which can be used in systems for search and rescue
(SAR) [11], blue force tracking (BFT), communication
intelligence (COMINT) systems [12], or local navigation
systems [13,14].
Equation (2) concerns the so-called two-dimensional SDF
method, considering the assumption that the signal emitted by
the source is a harmonic signal. Most works focused on the
SDF are based on this assumption. In free space, the DFS
estimation based on the received signal spectral analysis is
accurate and straightforward. In urban areas with multipath
propagation, the spectrum of the received signal called
Doppler spectrum is scattered, and the DFS estimation is
challenging. Nevertheless, in these conditions, the use of
statistical Doppler spectrum analysis in the SDF method
makes it possible to locate emission sources in an urbanized
environment. In the case of radio complex emissions (i.e.,
modulated signals), the application of the SDF requires
restoring the carrier frequency. The implementation of this
procedure is generally complicated. Only for discrete phase
modulations (e.g., M-ary phase-shift keying (PSK) or
quadrature amplitude modulation (QAM)), this procedure can
be performed in a relatively straightforward manner. In this
case, before frequency detection, the signal is raised to the
appropriate power. In the M-ary PSK, this power should be
equal to the M-order of the modulation. Thanks to this
procedure, modulation is eliminated, and the received signal
is a harmonic signal with at the M-times multiplied frequency
[15]. This allows us to reduce the detection band of the
instantaneous signal frequency to the range of
max
2D
f
. In our
solution, the Doppler shift frequency detection comes down to
determining the frequency for which the module of the
multiplied signal spectrum reaches its maximum value. In this
paper, the effectiveness assessment of the overlapping-based
procedure is presented based on the binary (BPSK), and
quadrature PSK (QPSK) signals recorded in real-environment
conditions.
III. OVERLAPPING PROCEDURE IN SDF METHOD
Figure 1 shows an example of a Doppler curve obtained
by practically performed measurements at the frequency
01832 MHzf=
[16]. This plot represents the DFS
relationship in the time (or distance) domain when considering
uniform receiver motion. In this case, in addition to the actual
measurement results (solid red line), we also showed a
theoretical curve for a given measurement scenario (dashed
blue line).
Fig. 1. Empirical (solid red line) and theoretical (dashed blue line) Doppler
curves for exemplary measurement results.
The reconstruction of the entire Doppler curve ensures
minimization of the emitter location errors, as it enables the
averaging of the results obtained based on (2) recorded along
the whole motion trajectory. From Fig. 1, we can see that the
empirical DFSs are fluctuating around the proper theoretical
values. Detection of the temporary values of
D
f
by means of
spectral analysis of the signal received in the time window
t
is one of the ways to reduce the impact of these fluctuations.
In this case, the detection comes down to determining the
frequency for which the spectral module obtained from the
received signal samples has a maximum value. However, this
method of minimizing fluctuations is strictly local in nature.
Therefore, the determination of
D
f
should be performed
based on the data recorded in the largest possible time interval
A
T
, in which the frequency change related to the Rx mobility
should be negligibly small. The overlapping method [8]
provides a solution to minimizing the location error based on
averaging the spectral analysis results. The proposed
procedure is depicted in Fig. 2 [16].
269 SPSympo 2021

Fig. 2. Diagram of overlapping procedure in SDF method.
In the illustrated scheme, we considered the parameters
resulting from the analyzed measurement scenario
(
01832 MHz,f=
60 km hv
).
Based on in-phased and quadrature (IQ) samples, the
spectral module of the recorded signal is determined in the
interval
100 ms.t=
An argument of its maximum value is
represented by
.
D
f
By repeating this procedure 40 times and
averaging the values obtained in the time window
4 s,
A
T=
the coordinates of the located object are determined. Then, the
time window of the analysis,
,
A
T
is shifted by the value of
A
t
(in Fig. 2,
 = 
A
tt
). We repeat the averaging procedure
and determining the emitter coordinates. In this way, we
obtain the possibility of both continuous assessment of the
located object position and minimization of the location error
by averaging all the results that we get along the entire Rx
trajectory.
IV. EFFECTIVENESS ASSESSMENT OF DEVELOPED METHOD
The effectiveness evaluation of the developed method
used comes down to the accuracy assessment for the Doppler
curve approximation. For this purpose, in relation to the
theoretical Doppler curve, root-mean-square errors (RMSEs)
are determined for the results obtained with and without the
overlapping method
For further analysis, we chose the PSK binary (BPSK)
signals recorded on the measurement route described in [15].
In this case, the signal was transmitted at the carrier frequency
01832 MHz,f=
and the receiver placed in the vehicle was
moving at a speed of about
37 km hv
. Figures 3 and 4
presents the sample Doppler curves without and with the
filtering gross errors, respectively. The empirical curve
( )
De
ft
was obtained based on the spectrum analysis of the
recorded signal. The DFS measurement points from this curve
are the basis for determining the approximating curve
( )
Da
ft
with a regular shape. The theoretical curve
( )
0D
ft
was
determined based on (1) for the real emitter coordinates. The
comparison of the appropriate DFS values (i.e., for the same
time moments) for the theoretical and approximating curves is
the basis for determining the RMSEs. In this case, we used the
empirical curve after the filtering process of the gross errors.
Fig. 3. Empirical (solid blue), approximation (dashed red), and theoretical
(dashed black) Doppler curves for exemplary recorded signal without
filtering gross errors.
Fig. 4. Empirical (blue cross points), approximation (dashed red), and
theoretical (dashed black) Doppler curves for exemplary recorded signal
with filtering gross errors.
In the effectiveness analysis of the overlapping method,
we assumed:
100 ms,=t
4 s,
A
T=
whereas
A
t
was
changing from 100 ms to 4 s with step 100 ms. The case for
=
AA
tT
corresponded to a non-overlapping approach.
Figures 5–7 show the obtained RMSE results versus
A
t
for three exemplary signals. When comparing the RMSEs for
different
,A
t
we considered the fact that the number of DFS
values used to determine individual errors was different.
Hence, the presented RMSEs consider this issue.
The obtained results clearly indicate that the application of
the overlay method significantly increases the accuracy of
DFS reconstruction, and therefore the accuracy of the
estimation of the location of the localized objects in the SDF
method. Increasing the
A
t
value contributes to increasing
the RMSE. On the other hand, the value of this parameter
should not be too low as it should consider the need for real-
time signal processing in the target location system.
270 SPSympo 2021

Fig. 5. RMSE of DFS for exemplary signal no. #1.
Fig. 6. RMSE of DFS for exemplary signal no. #2.
Fig. 7. RMSE of DFS for exemplary signal no. #3.
V. CONCLUSIONS
This paper focused on signal processing issues in the SDF
location technique. The developed solution, which involves
the introduction of the overlapping procedure, significantly
reduces the impact of fluctuations in the detected DFS values
on the accuracy of determining the localized-emitter
coordinates. The evaluation of the errors in the approximation
of the Doppler curve was obtained based on the measurement
data. This assessment showed the effectiveness of the applied
method of processing signals recorded in real conditions.
REFERENCES
[1] J. A. del Peral-Rosado, R. Raulefs, J. A. López-Salcedo, and G. Seco-
Granados, “Survey of cellular mobile radio localization methods: From
1G to 5G,” IEEE Commun. Surv. Tutor., vol. 20, no. 2, pp. 1124–1148,
Secondquarter 2018, doi: 10.1109/COMST.2017.2785181.
[2] J. F. Sanford, M. Potkonjak, and S. Slijepcevic, Localization in
wireless networks: Foundations and applications. New York, NY,
USA: Springer, 2012.
[3] A. Küpper, Location-based services: Fundamentals and operation.
Chichester, England; Hoboken, NJ, USA: Wiley, 2005.
[4] M. Srivatsa, A. Iyengar, J. Yin, and L. Liu, “Scalable key management
algorithms for location-based services,” IEEEACM Trans. Netw., vol.
17, no. 5, pp. 1399–1412, Oct. 2009, doi:
10.1109/TNET.2008.2010222.
[5] H. Abou-Zeid and H. S. Hassanein, “Toward green media delivery:
Location-aware opportunities and approaches,” IEEE Wirel. Commun.,
vol. 21, no. 4, pp. 38–46, Aug. 2014, doi:
10.1109/MWC.2014.6882294.
[6] R. R. Selmic, V. V. Phoha, and A. Serwadda, Wireless sensor
networks: Security, coverage, and localization. Springer International
Publishing, 2016.
[7] J. M. Kelner and C. Ziółkowski, “Effectiveness of mobile emitter
location by cooperative swarm of unmanned aerial vehicles in various
environmental conditions,” Sensors, vol. 20, no. 9, Art. no. 9, May
2020, doi: 10.3390/s20092575.
[8] M. H. H. Hayes, Digital signal processing, 2nd ed. New York, NY,
USA: McGraw-Hill Education, 2011.
[9] J. Rafa and C. Ziółkowski, “Influence of transmitter motion on
received signal parameters – Analysis of the Doppler effect,” Wave
Motion, vol. 45, no. 3, pp. 178–190, Jan. 2008, doi:
10.1016/j.wavemoti.2007.05.003.
[10] P. Gajewski, C. Ziółkowski, and J. M. Kelner, “Using SDF method for
simultaneous location of multiple radio transmitters,” in 2012 19th
International Conference on Microwave Radar and Wireless
Communications (MIKON), Warsaw, Poland, May 2012, vol. 2, pp.
634–637, doi: 10.1109/MIKON.2012.6233581.
[11] J. M. Kelner and C. Ziółkowski, “Portable beacon system for
emergency mountain landing pad,” in 2019 European Navigation
Conference (ENC), Warsaw, Poland, Apr. 2019, pp. 1–5, doi:
10.1109/EURONAV.2019.8714174.
[12] J. M. Kelner and C. Ziółkowski, “SDF technology in location and
navigation procedures: A survey of applications,” in Proceedings SPIE
10418, 2016 XI Conference on Reconnaissance and Electronic
Warfare Systems (CREWS), Ołtarzew, Poland, Apr. 2017, vol. 10418,
p. 104180B, doi: 10.1117/12.2269512.
[13] J. M. Kelner and C. Ziółkowski, “Doppler effect-based automatic
landing procedure for UAV in difficult access environments,” J. Adv.
Transp., vol. 2017, no. e8092718, pp. 1–9, 2017, doi:
10.1155/2017/8092718.
[14] C. Ziółkowski and J. M. Kelner, “Doppler-based navigation for mobile
protection system of strategic maritime facilities in GNSS jamming and
spoofing conditions,” IET Radar Sonar Navig., vol. 14, no. 4, pp. 643–
651, Apr. 2020, doi: 10.1049/iet-rsn.2019.0413.
[15] J. M. Kelner and C. Ziółkowski, “The use of SDF technology to BPSK
and QPSK emission sources’ location,” (in Polish), Przegląd
Elektrotechniczny, vol. 91, no. 3, pp. 61–65, Mar. 2015, doi:
10.15199/48.2015.03.14.
[16] R. Szczepanik, Integration of GPS receiver and digital map with
Doppler location system of radio emitters, Military University of
Technology, Warsaw, Poland, 2020.
271 SPSympo 2021