A Method of Localization with Multi-Channels UWB Bio-Radar in

Application of Through-Wall Detection

Ziqi Zhang, Fugui Qi, Yang Zhang, Hao Lv, Zhao Li, Miao Liu and Jianqi Wang

Department of Electronics, School of Biomedical Engineering, the Fourth Military Medical University, Xi’an 710032,

China.

Email: wangjq@fmmu.edu.cn

Keywords: Bbio-radar, multi-channels, localization, through-wall

Abstract: Recently, technology of ultra-wide band (UWB) bio-radar develops rapidly which becomes a very helpful

method of rescuing for post-disaster like earthquake and also useful surveillance medium for security fields

like anti-terrorists. There are two main aspects of UWB technology that researchers concern namely

technology of targets recognition and technology of targets localization. Most methods about targets

localization based on intersecting of multi arcs cannot determine the location of the targets effectively

because those arcs could not always intersect into one point. In this paper, we proposed a novel method

based on hyperbolic model using multi-channels bio-radar for two dimensional localization. Free-space and

through-wall experiments results indicated that the proposed method could determine the location of targets

accurately and effectively with an average error around 10cm between detection results and real positions.

1 INTRODUCTION

Ultra-wide Band (UWB) bio-radar is a novel kind of

radar combining radar technology and biomedical

engineering technology, and regards organism like

human bodies as the main detection targets, which is

one of the most effective technical methods to

search survivors that applied in some post-disaster

search and rescue operations like the earthquake, the

mine disaster, the debris flow and so on (Qi et al.,

2016; Lv et al., 2016; Liang et al., 2016). Bio-radar

could determine whether there is a life target and its

location within the detection area quickly, which

could provide guidance for professional rescue to

improve accuracy and efficiency of the rescue. The

detection results of bio-radar always contains two

aspects: targets recognition (whether there is a life

target and classification of target types), and targets

location. Actually, targets location technology is

eagerly needed in rescue operations, while it is still a

difficult problem to be solved at the moment

(Nguyen and Pyun, 2015; Monica and Ferrari, 2015).

If the actual position of the trapped target can be

accurately estimated, the limited search and rescue

resources will be better concentrated and the

efficiency of rescue will be greatly improved.

Currently, algorithm of multi arcs intersecting is

commonly used in two-dimensional positioning,

which can easily lead to a circumstance that

different arcs intersect into different points resulting

in failing to locate the life targets. Therefore, this

paper proposed a location method based on

hyperbolic model with multi-channels portable bio-

radar platform, and the experimental results of single

target location indicated that this method could

determine the two-dimensional coordinates of the

target more accurately. The location of multi-targets

is the most challengeable topic in location

technology for the absence of shadowing effect

(Kocur et al., 2011), research in this paper also lays

the foundation for solving the difficult problem of

multi-targets location.

This paper is organized as follow. In Section 2,

the multi-channels UWB bio-radar and antennas

array setup are introduced. Section 3 introduces the

new kind of two dimensional detection method for

life targets based on hyperbolic model and analyses

feasibility of it theoretically. The experimental

results are discussed in Section 4. The conclusion is

given in Section 5 finally.

Zhang, Z., Qi, F., Zhang, Y., Lv, H., Li, Z., Liu, M. and Wang, J.

A Method of Localization with Multi-Channels UWB Bio-Radar in Application of Through-Wall Detection.

In Proceedings of the International Workshop on Environment and Geoscience (IWEG 2018), pages 501-506

ISBN: 978-989-758-342-1

501

2 MULTI-CHANNELS UWB BIO-

RADAR SYSTEM AND

EXPERIMENTAL SETUP

2.1 Multi-Channels UWB Bio-Radar

System

As Figure 1 shows, The multi-channels portable

UWB bio-radar system we adopt consists of UWB

bio-radar antennas with 1 transmitting antenna and 4

receiving antennas, a portable multi degree of

freedom foldable support and a touch-sensitive

wireless processor. Its operating central frequency of

the UWB radar is 400MHz with a band width of

400MHz. The pulse repetition rate is 128 KHz, and

the AD sampling frequency for each channel is 64

Hz, which is sufficient to capture the instantaneous

change of human motion.

Figure 1: Multi-channels UWB bio-radar system.

2.2 Antennas Array and Experiment

Setup

According to previous research (Zhang et al., 2006),

it needs at least 3 channels to avoid “ghost” if the

two dimensional location for a target should be

determined. As Figure 2 shows, an antenna

combination of one transmitting antenna and three

receiving antennas (receiving antennas are marked

as R

, R

and T for transmitting antenna) is

adopted in this paper. The transmitting antenna is

placed in the middle of the array together with the

receiving antenna R

on the top of it, another two

receiving antennas are on the two sides of antenna T

with a distance of 1.08m (the folding arms of the R

and R

are extended to the longest) repectively to

ensure the angle resolution when targets locate at a

relatively long distance. Therefore, they show a

linear array as Figure 2. Thereinto, the box T

contains not only transmitting antenna but host

which controls wifi module connects with wireless

peocessor and coordinate the linear antenna array

smoothly.

Figure 2: A schematic diagram of a linear antenna array.

In this paper, we intend to verify the

effectiveness of the new location method via two

different experimental scenarios, namely free-space

and through-wall. Firstly we conduct the free-space

experiment to verify the principle and then carry out

the through-wall detection experiment, so as to

gradually verify the rationality and correctness of

the novel location method. The two experimental

scenes are illustrated by Figure 3 to Figure 6.

Figure 3: Antenna deployment in free-space scenario.

Figure 4: Experimental scene in free-space scenario.

Wireless processor

Antennas

Foldable suppor

IWEG 2018 - International Workshop on Environment and Geoscience

502

Figure 5: Antenna deployment in through-wall scenario.

Figure 6: Experimental scene in through-wall scenario.

3 TWO DIMENSIONAL

LOCALIZATION METHOD

WITH MULTI-CHANNELS BIO-

RADAR

The principle of the localization method is shown in

Figure 7, the antennas array seems like a line from

high angle, the receiver R

and transmitter T are

approximately treated as the same point. In Figure 7,

the traveling time of microwave from T to target P is

denoted as τ

, while τ

,τ

and τ

denote the traveling

time of reflected microwave from target P to R

, R

and R

respectively.

Then, some distances are calculated by

TP 





(1)

PR 





(2)

PR 







(3)

PR 









(4)

where c is the speed of light.

The distance of microwaves that sent from the

transmitter then reflected by the target P to receiving

antennas R

, R

, and R

are L

, L

, and L

respectively, then we can get









02 0 2

L L TP PR TP PR

PR PR

   





   

 

(5)

Therefore, the absolute value of difference

between



与



is equal to that between L

and

which could be acquired by bio-radar easily.

According to the knowledge of analytic geometry, if

the absolute value of the distance difference between

the point P and another two points on the plane is

fixed, then the point P is located on one branch of

the hyperbola with the focus of R

and R

Mathematically,

h2 0

cRTTR



(6)

02 0 2

aLL PR PR

 



 

(7)

Among them, c

is a half focal length of

hyperbola, a indicates the real half axis of hyperbola,

so the imaginary half axis of hyperbola is calculated

bca



. Taking a set of data in the

experiment (whose actual coordinates are (-1,-5)) for

an example, then the hyperbola determined by R

and T is acquired as Figure 8 shows. If L

, the

target is in the left branch of the hyperbola in the

third quadrant of Cartesian coordinates; If L

, the

target is in the right branch of the hyperbola in the

fourth quadrant of Cartesian coordinates. While, it is

not adequately to calculate on which specific point

of that branch the target locates. Besides, we treat T

and R

as the same point, so we can suppose that

=τ

, namely

TP PR



 

. Then target P is also on the

circle with the radium of r as shown in Figure 9.

TP PR







   

(8)

A Method of Localization with Multi-Channels UWB Bio-Radar in Application of Through-Wall Detection

503

(8)

If we plot that hyperbola and arc described above

in the same Cartesian coordinate system, they will

intersect at two points as shown in Figure 10. For

this sample of data, due to the L

, we get (-1.056,

-5.077) by tracking the intersection of the two

curves in the third quadrant. Thus, we accomplished

the two-dimensional positioning of the target P. In

addition, the positioning result have little error

compared with actual position of target namely (-1, -

5). The absolute error is less than 8cm in this

example.

Figure 7: The principle of the localization method.

Figure 8: Hyperbola determined by R

, R

and T.

Figure 9: Circle determined by R

and T.

Figure 10: The method of the ensure the specific

coordinates of the target.

4 EXPERIMENT RESULTS AND

DISCUSSION

In order to describe the position of the target more

easily, we have set up a Cartesian coordinate system

on the ground of the laboratory as Figure 11 and

Figure 12 show. The target positions were divided

into 25 different points and marked with

corresponding number. The position of the host of

bio-radar is set as the origin of coordinates, both

receiving antennas R

and R

are lied in horizontal

axis. In free-space experiment, most coordinates of

position are integer value except for positions on

right sides of the coordinate system because of the

limitation of the experimental site. In through-wall

T(R

) R

1.08m

1 08m

IWEG 2018 - International Workshop on Environment and Geoscience

504

experiment, we let the bio-radar system turn around

and carried out the experiment in opposite direction,

so positions on left sides of the coordinate system is

not integer value as shown in figures below.

Figure 11: Target position division of free-space

experiment.

Figure 12: Target position division of through-wall

experiment.

We calculate the errors between detection results

and real positions. As for the example referred

above, the detection results is (-1.056,-5.077) and

the real coordinates of the positions is (-1,-5), so the

distance between (-1.056,-5.077) and (-1,-5) namely

error is 0.095m. To cooperate with the Cartesian

coordinate system set above, experimental results

are listed in Table1 and Table 2 in the form of

matrix below. Among them, color red indicates a

lager error but color green means a smaller error.

What’s more, we drew data of both the two

scenarios into one line diagram as Figure 13 shows.

Table 1: Errors between detection results and real

positions of free-space

Ordinate

Abscissa

-2 -1 0 1 1.7

-2 0.122 0.014 0.090 0.143 0.098

-3 0.071 0.082 0.090 0.020 0.152

-4

0.134 0.130 0.060 0.158 0.092

-5

0.161 0.095 0.100 0.161 0.122

-6

0.192 0.221 0.050 0.142 0.067

Unit of data in the table: m

Table 2: Errors between detection results and real

positions of through-wall

Ordinate

Abscissa

-1.7 -1 0 1 2

-2

0.175 0.277 0.210 0.036 0.121

-3

0.054 0.071 0.140 0.102 0.194

-4

0.153 0.100 0.160 0.140 0.028

-5

0.122 0.114 0.120 0.126 0.100

-6

0.291 0.110 0.130 0.014 0.106

Unit of data in the table: m

Figure 13: Errors between detection results and real

positions of two scenarios.

5 CONCLUSIONS

This paper proposed a novel two-dimensional

localization method for life target detection using

multi-channels bio-radar based on hyperbolic model.

Actual experimental results show that the proposed

method could avoid the circumstance that different

A Method of Localization with Multi-Channels UWB Bio-Radar in Application of Through-Wall Detection

505

arcs fail to intersect into one point. Moreover, it

could accurately determine the coordinates of a life

target no matter in application of free-space or

trough-wall scenarios.

ACKNOWLEDGMENTS

Research in this paper was supported by National

Natural Science Foundation of China (Grant No.

61327805), Shaanxi Technology Committee (Grant

No. 2016KJXX-03) and National Natural Science

Foundation of China (Grant No. 31600796).

REFERENCES

Kocur D, J Rovňáková and D Urdzík 2011 Mutual

shadowing effect of people tracked by the short-range

UWB radar International Conference on

Telecommunications and Signal Processing

Liang F, et al. 2016 Detection of Multiple Stationary

Humans Using UWB MIMO Radar Sensors 16(11)

1922

Lv H, et al. 2016 Improved Detection of Human

Respiration Using Data Fusion Based on a Multistatic

UWB Radar Remote Sensing 8(9) 773

Monica S and G Ferrari 2015 UWB-based localization in

large indoor scenarios: optimized placement of anchor

nodes IEEE Transactions on Aerospace & Electronic

Systems 51(2) 987-999

Nguyen V H and J Y Pyun 2015 Location Detection and

Tracking of Moving Targets by a 2D IR-UWB Radar

System Sensors 15(3) 6740-6762

Qi F, et al. 2016 Detection and Classification of Finer-

Grained Human Activities Based on Stepped-

Frequency Continuous-Wave Through-Wall Radar

Sensors 16(6) 885

Zhang Y, et al. 2006 Secure localization and

authentication in ultra-wideband sensor networks

IEEE Journal on Selected Areas in Communications

24(4) 829-835

IWEG 2018 - International Workshop on Environment and Geoscience

506