close

Вход

Забыли?

вход по аккаунту

?

Frequency-detected acoustic ranging solutions in wireless sensor networks an experimental study.

код для вставкиСкачать
ASIA-PACIFIC JOURNAL OF CHEMICAL ENGINEERING
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
Published online 17 October 2008 in Wiley InterScience
(www.interscience.wiley.com) DOI:10.1002/apj.205
Special Theme Research Article
Frequency-detected acoustic ranging solutions in wireless
sensor networks: an experimental study
Jiming Chen,1 * Xingfa Shen,2,1 Huijin Ren1 and Youxian Sun1
1
2
State Key Lab of Industrial Control Technology, Institute of Industrial Process Control, Zhejiang University, Hangzhou 310027, P. R. China
College of Computer, Hangdian University, Hangzhou, P. R. China
Received 20 April 2007; Revised 13 July 2008; Accepted 25 July 2008
ABSTRACT: Ranging is a basic distance estimation for many range-based localization approaches, which are important
to wireless sensor networks (WSNs) applications at many levels. In this paper, we propose a simple frequency-detected
based time difference of arrival (FD-TDoA), which can be implemented by detecting acoustic frequency to compute
the time of flight in air. Furthermore, we put forward a new acoustic ranging solution named time of arrival (TOA)2 ,
which can be applied to WSNs with asynchronous nodes. Unlike other published works, the design of TOA2 uses a
bidirectional acoustic signal exchange between a pair of communication nodes. This technique is significantly simple
and effective. The latency between the time at which the Mica2 is commanded to emit an acoustic pulse and the earliest
possible time that can be detected anywhere, is considered in our solutions. The error of these ranging solutions, the
correction expressions by fitted lines and the sensitivity on hardware are analyzed by a large number of experiments
based on a resource-constrained Mica2 hardware platform.  2008 Curtin University of Technology and John Wiley
& Sons, Ltd.
KEYWORDS: acoustic signal detection; ranging; sensor networks; asynchronous
INTRODUCTION
Wireless sensor networks (WSNs) have been gaining
great attention among the research community in the
past few years, because the recent developments in
Micro-electro-mechanism system (MEMS) technology
have provided us with a cheap, customizable, embedded sensor system capable of wireless communication.
The fields of their possible applications vary from
military target tracking to habitation monitoring.[8] A
fine-grained geographic localization of nodes that provides high-precision location information, is essential
for these potential distributed sensor networks applications, especially in systems of positioning, geographic
location-aware routing, collaborative sensing, and target tracking, etc.[8] . For example, in a system of mobile
robots, which is controlled in a distributed intelligent
sensor network, localization is the principal problem
that should be investigated.[11]
Localization in sensor networks is most commonly
range-based and accomplished using range estimation among sensor nodes, except for some particular
range-free localization approaches.[6][13] The range-free
*Correspondence to: Jiming Chen, State Key Lab of Industrial
Control Technology, Institute of Industrial Process Control, Zhejiang
University Hangzhou 310027, P. R. China. E-mail: jmchen@ieee.org
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
localization approaches make no assumption about the
availability or validity of distance information, but the
range-based localization should have knowledge about
the distances between the nodes.
A large amount of work has been done for ranging
techniques in the past few years, which calculate
absolute point-to-point distance. Among them, the most
classical technique is to measure the time of arrival
(TOA) of signals through air. TOA-based approaches,
such as GPS to measure pure RF time of flight, have
many limitations when applied to WSNs.[18] There
are a number of problems which make measuring RF
time of flight quite difficult in a low-cost system. In
order to measure time of flight, precise measurement
is demanded and highly tight synchronization must be
kept between the sender and the receiver, which cannot
be implemented on a low-power, resource-constrained
hardware platform like Mica2 Motes.
So, researchers consider the approaches using an
acoustic signal that has many advantages over RF-based
approaches. Measuring acoustic time of flight is much
easier because the sound propagates much slower than
the radio signals in air, and their synchronization can
easily be achieved using RF. One of the most successful
and popular approaches is called time difference of
arrival (TDoA), which can estimate the time difference
of flight of acoustic and radio signals.[9][10][15] Similar
590
J. CHEN ET AL.
designs such as the Active Bat rely on instrumenting
the environment with carefully calibrated sensors.[17]
Furthermore, an angle of arrival (AOA) technique has
been proposed that allows the nodes to estimate and map
relative angles between neighbors.[16] Received Signal
Strength Indicator (RSSI) technology such as RADAR
has been proposed for resource-constrained systems,
which translates the signal strength into distance estimation by theoretical and empirical models.[7][14]
In fact, there also exist some works on acoustic ranging or localization. In,[12] Mungamuru et al . described
an approach to estimate the source orientation as a consequence of the proposed sound localization technique
by taking into account the source directivity, microphone directivity, and source-microphone distances.
Chen et al . studied acoustic ranging to determine the
position of a seafloor transponder, but with significant
uncertainties inherent in GPS measurements and the use
of a high-cost commercial acoustic transponder.[2] In,[1]
they presented a localization algorithm to self-organize
a global coordinating system on an ad hoc WSN, which
works well combined with acoustic ranging in practice.
In this paper, we propose a simple frequency-detected
based approach for TDoA (FD-TDoA), which records
the arrival time point by detecting the frequency of
acoustic signal unlike existent TDoA approaches, which
use an acoustic signal peak to estimate time of flight.
The latency between the time at which the sensor node
is commanded to emit an acoustic pulse and the earliest possible time it can be detected anywhere can
be compensated by frequency self-detection. Furthermore, we present a new acousticranging solution which
can be applied to WSNs with asynchronous nodes.
The approach is named as TOA2 that measures TOA
of acoustic signal in asynchronous sensor nodes. This
designed solution exchanges acoustic signals between a
pair of communication nodes and easily gets the time of
flight, which has low resource requirements. It should be
mentioned that the idea of message exchanging has been
applied to synchronizing a pair of nodes, but has not
been introduced to ranging estimation.[5] These solutions have been implemented in a resource-constrained
standard Mica2.
The remainder of the paper is organized as follows. The section on Frequency-Detected Based TDoA
introduces our proposed FD-TDoA and describes the
self-detected approach in detail. In the section on
Acoustic Ranging Solution TOA2 in Asynchronous
Mode, the design and deduction process of TOA2
are analyzed. The section on Experimental Evaluations follows with a more detailed introduction of standard Mica2 Motes experimental hardware platform and
presents initial results of experiments. The final section
offers some concluding remarks and ongoing works.
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
Asia-Pacific Journal of Chemical Engineering
FREQUENCY-DETECTED BASED TDOA
The most important content of ranging solutions is
to measure the time of flight of the acoustic signals
between sender and receiver. In order to get an accurate
time of flight, a sophisticated synchronization mechanism on a pair of communication nodes is essential.
TDoA uses radio signals to notify sensor nodes to realize synchronization. Because the propagation speed of
radio signals is 106 times higher than the speed of
sound, the differences of arrival time of the acoustic
and radio signals can be used to estimate the time of
flight of sound propagation.
As shown in Fig. 1, both node ‘a’ (Na ) and node ‘b’
(Nb ) are a communication node-pair. Assume T as
the absolute time difference of system clock between
asynchronous Na and Nb , so T + Ta = Tb , where,
Ta and Tb are defined as the current time of system
clock in Na and Nb respectively. t1 , t3 are the time
points that radio and acoustic signals are emitted
at Na , and they are received at t4 and t8 at Nb
correspondingly. t5 is the time point that acoustic signal
is self-detected by the microphone at Na . t2 equaling
to t1 + T is the corresponding time point of t1 at
Nb .
The error of introducing radio signals to make the
communication node-pair synchronous includes two
parts: modulation and travel in air of the radio signal.
The latter is very small because of the high propagation
speed of radio signals, but the modulation time cannot
be neglected. In Fig. 1, this error can be described by
1 = t4 − t2 .
In general, it is thought that radio and acoustic signals
are emitted simultaneously. But in our experiment
designed on a resource-constrained distributed sensor
network hardware platform, in order to eliminate the
effect of RF on acoustic signals, we set a constant delay
2 between t1 and t3 before starting up the sounder.
But the most difficult aspect in FD-TDoA approach
is to compute the latency latency between t3 and the
time point t5 indicating the exact time point that the
acoustic signal is produced on the side of the sender
node (as shown in Fig. 1). Obviously, latency should
be considered in measuring the time difference.
Na
t1
∆2
∆latency
Nb
t2
t4
∆1
t3
t6
t5
∆4
t8
Figure 1. Time difference of arrival approach.
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
Asia-Pacific Journal of Chemical Engineering
FREQUENCY-DETECTED ACOUSTIC RANGING SOLUTIONS IN WSNs
In general, latency in question is unknown and a
little difficult to be obtained, because the Motes system
does not provide the information of the sounder. In
the FD-TDoA approach, when TinyOS system calls
Sounder.StdControl.start to emit the acoustic signal, this
time point is recorded by function SysTime.getTime32(),
which corresponds to t3 . A matching sensor microphone
in the sensor board of MTS310CA is started up and the
system calls the function Mic.readToneDetector(). At
the sender Na detects the distinct sound (which we call
a ‘chirp’) that is produced at the sender nodes.
In order to avoid disturbance of sudden noise with
the same frequency, it is confirmed that a ‘chirp’ is
detected when function Mic.readToneDetector() returns
n-continual 0 (the value n is set by the operator
in the program according to the environment). The
microphone will not detect the acoustic signal and the
ranging experiment will fail if the duration of the noise
with the same frequency is larger than n × p, where p
is sampling period of the microphone.
Our approach is differentiated from existing TDoA
approaches by starting up a microphone at the sender
node, which we call self-detected. Because the acoustic
signals are of a fixed frequency with slight variations
of 4.5 kHz, function Mic.readToneDetector() detecting
this frequency will record the time point using function
SysTime.getTime32(), which corresponds to t5 . So the
latency latency can be measured accurately in our
experiment. t8 can be obtained by using a microphone
to detect the frequency at the Nb .
Define tFD−TDoA as the time difference of acoustic and
radio signals that fly from Na to Nb . Because 1 cannot
be obtained, it generally is negligible in many TDoA
approaches. So Eqn (1) is generally used to estimate the
time of flight. If a known value for the speed of sound
propagation V in air is given, the distance dFD−TDoA
traveled by the waves can be computed by Eqn (2)
tFD−TDoA = 3 = t8 − t6 = t8 − (t2 + 2 )
= t8 − t2 − 2
dFD−TDoA = 3 × V
(2)
= 3 − latency = (t8 − t6 ) − (t5 − t3 )
= (t8 − t2 − 2 ) − (t5 − t3 )
dFD−TDoA
= tFD−TDoA
×V
ACOUSTIC RANGING SOLUTION TOA2 IN
ASYNCHRONOUS MODE
The classical TOA ranging solutions need to make the
clock synchronous firstly between the communication
node-pair. The FD-TDoA also introduces a radio signal
to achieve the function of clock synchronization, which
adds extra error 1 to the ranging estimation.
In order to provide a more simple and effective
ranging solution, we propose a novel idea, which can be
applied to the asynchronous communication node-pair
and differs from previous classical TOA approaches. It
should be stated that the idea was deduced from[5] that
has been applied to make the node-pair synchronization.
A similar work has been done in,[4] but our solution to
estimate distance is frequency-detected, which is more
simple and can be implemented on low-power resourceconstrained hardware.
As displayed in Fig. 2, a ‘chirp’ will be emitted
from Na and received at Nb . The time points ta in Na
are recorded by the function SysTime.getTime32() when
system calls Sounder.StdControl.start. The program in
Nb calls Mic.readToneDetector() to detect the frequency
of ‘chirp’ and then records the time tb .
In the same way, we record the time point of emitting
tb and receiving ta acoustic signals at Nb and Na
respectively. The time of flight tTOA2 will be calculated
easily. According to our foregoing assumption, we can
get Eqn (5) from Subfig A in Fig. 2.
ta + T + tTOA2 = tb
(5)
Based on Subfig B in Fig. 2, another equation Eqn (6)
can be given when Na and Nb exchange their roles.
tb − T + tTOA2 = ta
(1)
In practical experiments, we can re-estimate the differ
and distance dFD−TDoA
by taking
ence time tFD−TDoA
latency into account using the following Eqn (3) and
Eqn (4).
tFD−TDoA
sender and receiver can be estimated easily by Eqn (2)
and Eqn (4).
(6)
From Eqn (5) and Eqn (6), the time of flight of acoustic
signals tTOA2 can be deduced from the following Eqn (7)
theoretically:
Na
ta′
Nb
ta′ + ∆T
Na
tb′′ − ∆T
t TOA
′ 2
(3)
Nb
tb′′
t TOA
′ 2
tb′
ta′′
(4)
SubFig A
Based on Eqn (3), only time points t2 , t3 , t5 , t8 should
be recorded, and the point-to-point distance between
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
SubFig B
Figure 2. Acoustic ranging approach TOA2 .
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
591
592
J. CHEN ET AL.
Asia-Pacific Journal of Chemical Engineering
tTOA2 =
ta + tb − ta − tb
2
(7)
In our TOA2 approach, only the time points of ta ,
tb , ta , tb are recorded. It is not necessary to make
the communication node-pair synchronous, which can
avoid complex steps of synchronization and reduce
unnecessary errors caused by synchronization. The time
of flight of acoustic signals can be computed accurately
in theory and the distance dTOA2 can be estimated well
with the assumption of known sound speed by Eqn (8).
dTOA2 = tTOA2 × V
(8)
But the proposed approach will be implemented in
the resource-constrained Mica2 Mote system, which
is different from[15] and[17] that need high cost and
an expensive hardware platform. Therefore, the same
problem should be to deal with the latency between
the time at which the mote is commanded to emit
an acoustic pulse and the earliest possible time it can
be detected anywhere. We also adopt the manner of
acoustic signals self-detected to measure the latency at
sender nodes, which has been described in detail in FDTDoA design in the previous section.
Considering the error value t1 , t2 as shown in
Fig. 3, because the communication between Na and Nb
is peer-to-peer, it can be thought t1 = t2 = latency .
The above equations should be corrected by taking
latency into account.
+ T + tTOA
ta1
2 + latency = tb1
(9)
− T + tTOA
tb1
2 + latency = ta1
(10)
So the expression for the corrected time of flight of
acoustic signals tTOA
2 can be obtained in the Eqn (11)
and the distance is estimated by TOA2 with selfdetected by Eqn (12).
tTOA
2 =
ta1
+ tb1
− ta1
− tb1
− latency
2
×V
dTOA
2 = t
TOA2
Na
Nb
t a′ 1
∆ t1
t a′ 2
Na
ta′ 1 + ∆T
Nb
t ′′
∆ t2 b 1
t b′′2
t b′′1 − ∆T
t TOA
′ 2
t TOA
′ 2
t b′ 1
t a′′1
SubFig A
SubFig B
Figure 3. TOA2 acoustic ranging approach based Mica2
Motes.
The performance of the approach will be evaluated on
the Mica2 Motes hardware test-bed in the next section.
EXPERIMENTAL EVALUATIONS
Experimental platform
Our FD acoustic ranging solution FD-TDoA and TOA2
approaches are tested on the Mica2 Motes system,
which is one of the most popular and commercially
available redeveloped research experimental platform
for typical low-power resource-constrained WSNs marketed by CrossBow technologies.[3] Applications for
Mica2 Motes are developed on an operating system
called TinyOS. TinyOS is an Event-based OS and its
application program is written using a language called
Nested C (NesC). NesC is a component-based structured language. It is an extension to C programing language and mainly used in embedded networks systems.
The device features an 8-bit Atmega Micro-controller
(Atmega 128L) with 4 Kb system RAM, 128 Kb
flash program memory, 8 channel, 10-bit ADC and
4 hardware timers. For I/O it has one external UART,
one SPI port and 53 general purpose I/O lines. It has a
cc1000 coprocessor with high sensitivity for wireless
communication, and its transceiver supports transfer
rates up to 76.8 Kbps with an increased radio range of
500 feet. There is an expansion connector I/O system
interface which allows a variety of sensing boards.
Among them, the microphone and the fixed-frequency
sounder are utilized by the acoustic ranging application
in this paper.
(11)
Original results of FD-TDoA and TOA2
(12)
In this approach, it can be found that the measured
time points can easily be recorded in the Mica2 Motes
experimental platform. Compared with TOA or TDoA
that needs synchronization, TOA2 looks more simple.
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
We measure the time of flight of acoustic signals at three
different scenarios (indoors, outdoors, in the aisle). A
pair of communication nodes are placed 50 cm above
the ground. The measured distance varies from 0 to
1000 cm with an interval 50 cm.
Ranging experiments at each measured point are
repeated 15 times including failed ones which will
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
Asia-Pacific Journal of Chemical Engineering
FREQUENCY-DETECTED ACOUSTIC RANGING SOLUTIONS IN WSNs
return a distance value of zero. Before starting each time
of the experiment, the temperature should be recorded
in advance to correct the speed of sound propagation
according to Eqn (13).
8
7
6
(13)
m
V = (331.45 + 0.59T )
9
T( ◦ C) is scaled as centigrade temperature and V (m/s)
is the T -based speed.
The original ranging estimation results of FD-TDoA,
TOA2 , which do not consider self-detected correction,
are plotted in Figs. 4–6 for the corresponding three
different scenarios: indoors, outdoors, in the aisle. The
data in these figures are estimated from both Eqns (2)
and (8). It is found that the error will be much larger
with measured distance.
In the acoustic ranging experiments, there are many
detection failures when measured distance exceeds
an upper limit, and that it cannot be scaled. It is
9
8
7
m
6
5
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
0
1
2
3
4
m
5
6
7
8
Figure 4. Original results of FD-TDoA, TOA2
indoors; T = 18 ◦ C.
9
8
7
m
6
5
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
0
1
2
3
4
m
5
6
7
8
Figure 5. Original results of FD-TDoA, TOA2
outdoors; T = 22 ◦ C.
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
5
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
2
mean value of TOA
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
0
1
2
3
4
m
5
6
7
8
Figure 6. Original results of FD-TDoA, TOA2 in the
aisle; T = 13 ◦ C.
the main disadvantage of almost all acoustic ranging
solutions, which have limited effective distance. In the
experiments, because of resource-constrained hardware
platform, the nodes cannot detect the ‘chirp’ when the
distance exceeds 600 cm outdoors, 1000 cm indoors
or 700 cm in the aisle. The measured data are thought
to be invalid if the absolute error is greater than 20
cm. So in many cases, even if the node can hear the
acoustic signals of each other, the error may still be
beyond 20 cm and cannot be accepted. According to
our experimental data, the effective distances for FDTDoA and TOA2 are 700 cm and 500 cm indoors, 600
cm and 550 cm outdoors, 600 cm and 550 cm in the
aisle.
The broken and bold lines are fitted based on the
mean value of measured points in an effective distance for FD-TDoA and TOA2 respectively. TOA2 has
smaller error than FD-TDoA but larger slope of fitted
line in the effective distance. In the experiment, it is
noticed that the times-of-failure measurement increases
with measured distance. It should also be mentioned
that there is a grey zone in the ranging. For example,
TOA2 cannot always get an effective measured value at
300 cm outdoors.
It is shown that measured results always are larger
than the actual value, which also illustrates that the
latency between command given and signal detected
really exists.
Corrected results of FD-TDoA and TOA2
As stated in Fig. 1 and Fig. 3, there is a latency latency .
It is necessary to get the latency by experiments. The
mean value of latency equals to 2 ms using a selfdetected approach by repetitious experiments. So the
above results can be improved by taking latency latency
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
593
J. CHEN ET AL.
Asia-Pacific Journal of Chemical Engineering
into account. According to Eqn (4) and Eqn (12), the
new improved results are plotted in Fig. 7, Fig. 8 and
Fig. 9 respectively. The expressions by the fitted lines
after correction are listed in the Table 1.
For the mean values of these measured data, the final
output results should be modified by these corrected
expressions of fitted lines. The final output results are
shown in Fig. 10 and Fig. 11.
From the Fig. 10 after correction, it is illustrated that
the results of FD-TDoA are evidently improved. In the
9
8
7
6
5
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
9
0
1
2
3
4
5
6
7
8
8
Figure 9. Corrected results of FD-TDoA, TOA2
in the aisle; T = 13 ◦ C.
7
m
6
5
8
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
7
6
5
1
2
3
4
m
5
6
7
m
4
0
8
3
2
Figure 7. Corrected results of FD-TDoA, TOA2
indoors; T = 18 ◦ C.
outdoor
indoor
in the aisle
ideal line
1
0
−1
9
0
1
2
3
4
5
6
7
m
8
Corrected output value of FDTDoA. This figure is available in colour online
at www.apjChemEng.com.
Figure 10.
7
6
m
594
5
4
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
fitted line for FD–TDoA
fitted line for TOA2
ideal line
3
2
1
0
0
1
2
3
4
m
5
6
7
8
Figure 8. Corrected results of FD-TDoA, TOA2
outdoors; T = 22 ◦ C.
three scenarios, the maximum absolute error is 30 cm at
point 300 cm outdoors and the minimum absolute error
can be reached 1 cm at point 350 cm indoor and 150
cm in the aisle, but almost all error rates are below 5%.
In Fig. 10, we can find that FD-TDoA is not sensitive
to environment.
Figure 11 demonstrates that TOA2 also gets a corrected improvement. Even if there are several singular values, the corrected expression still is effective to
restrain the error caused by a large slope especially at
long measured distance.
Table 1. Corrected expressions.
FD-TDoA
Outdoors
Indoors
In the aisle
x=
x=
x=
y−0.2224
1.1743
y−0.3741
1.1407
y−0.5408
1.1419
TOA2
x=
x=
x=
y+0.0882
1.2202
y+0.0847
1.2252
y+0.1419
1.2344
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
Sensitivity of hardware on results
In order to investigate the effect of hardware on the
ranging solutions, we use three different Mica2 communication node-pairs in the acoustic ranging experiments.
The results are plotted in the Fig. 12. The maximum
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
Asia-Pacific Journal of Chemical Engineering
FREQUENCY-DETECTED ACOUSTIC RANGING SOLUTIONS IN WSNs
platform. But sensor networks nodes like the Mica2
Motes hardware platform are resource-constrained, and
the characteristic of hardware like latency of sounder
should still be taken into consideration.
Our future works will try to develop the ultrasonic
signal-based ranging solution for these approaches. The
topic of range measurement over multiple pairs and
multiple hops will be addressed. Of course, applying our
acoustic ranging approaches to range-based localization
based on a mobile target tracking system also is a
possible work.
7
6
5
m
4
3
outdoor
indoor
in the aisle
ideal line
2
1
0
0
1
2
3
m
4
5
6
Acknowledgement
Figure 11. Corrected output value of TOA2 .
This figure is available in colour online at
www.apjChemEng.com.
4
3.5
node
9, 14
3
m
2.5
2
node
4, 10
This work is supported by NSFC-Guangdong Province
Union Project (No. U0735003); NSFC (No. 60604029);
NSF of Zhejiang (No. Y106384, Y107309); the Science and Technology Project of Zhejiang Province
(No. 2007C31038), the Specialized Research Fund
for the Doctoral Program of Higher Education (No.
20050335020),
and
863
High-tech
Project
(No. 2007AA041201).
node
11, 15
1.5
1
0.5
0
original measurements of FD–TDoA
original measurements of TOA2
mean value of FD–TDoA
mean value of TOA2
Figure 12. Results under different node-pair in the aisle.
measured difference of these node-pairs is 14 cm for
FD-TDoA and 6 cm for TOA2 , which indicates that the
result is insensitive to the particular node-pair and can
be reproduced for different instances of the hardware.
CONCLUSION
In this paper, we proposed acoustic ranging approaches
FD-TDoA and TOA2 . Error caused by latency of
sounder is the most serious problem to solve in finegrained acoustic ranging. It is important to estimate the
error, irrespective of whether FD-TDoA or TOA2 is
applied. We use the microphone in the sensor board of
Mica2 Motes to self-detect the acoustic signals in the
sender node. The average value of measurement is used
to approximate the latency of the sounder. Performance
of FD-TDoA and TOA2 can be improved significantly
by taking account of latency error.
TOA2 approach keeps away from the problem of
synchronization and simplifies the steps of acoustic
ranging. In theory, TOA2 can achieve the absolutely
accurate result with the support of a stronger hardware
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
REFERENCES
[1] J. Bachrach, R. Nagpal, M. Salib, H. Shrobe. Telecomm. Syst,
2004; 26(2–4), 213–233.
[2] H.H. Chen, C.C. Wang. Ocean Eng., 2007; 34, 2385–2399.
[3] Inc.Crossbow Technology. http://www.xbow.com/support/
support pdf files/Product Feature Reference Chart.pdf.
[4] B.C. Dalton, V.M. Bove. Audio-based self-localization for
ubiquitous sensor networks. In Proceedings of 118th Audio
Engineering Society Convention, Barcelona, Spain, 2005.
[5] S. Ganeriwal, R. Kumar, M.B. Srivastava. Timing-sync
protocol for sensor networks. In Proceedings of the 1st
International Conference on Embedded Networked Sensor
Systems, Los Angeles, Nov. 2003; 138–149.
[6] T. He, C. Huang, B.M. Blum, J.A. Stankovic, T.F. Abdelzaher. Range-free localization schemes for large scale sensor
networks. In Proceedings of ACM/IEEE International Conference on Mobile Computing and Networking (MOBICOM),
San Diego, California, USA, 2003; 81–95.
[7] J. Hightower, R. Want, G. Borriello. SpotON: An indoor
3D location sensing technology based on RF signal
strength. Technical Report UW CSE 00-02-02. University
of Washington, Department of Computer Science and
Engineering, Seattle, February 2000.
[8] I.F. Akyildiz, T. Melodia, K.R. Chowdury. IEEE Wireless
Commun., 2007; 14(6), 32–39.
[9] S. János, B. György, M. Miklós, L. Ákos, K. Branislav.
Acoustic ranging in resource-constrained sensor networks.
In Proceedings of The International Conference on Wireless
Networks, Las Vegas, Nevada, USA, 2004; 467–473.
[10] G. Lewis. Development and characterization of an acoustic
rangefinder. Technical Report 00-728, University of Southern
California, Department of Computer Science, Los Angeles,
2000.
[11] K. Morioka, J.H. Lee, H. Hashimoto. IEEE Trans. Ind.
Electron., 2004; 51(1), 229–237.
[12] B. Mungamuru, P. Aarabi. IEEE Trans. Syst. Man Cybern. Part
B-Cybern., 2004; 34(3), 1526–1540.
[13] D. Niculescu, B. Nath. Telecomm. Syst., 2003; 22(1–4),
267–280.
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
595
596
J. CHEN ET AL.
[14] P. Bahl, V.N. Padmanabhan. Radar:an in-building rf-based
user location and tracking system. In Proceedings of the IEEE
INFOCOM, Tel-Aviv, Israel, 2000; 775–784.
[15] N.B. Priyantha, A. Chakraborty, H. Balakrishnan. The cricket
location-support system. In Proceedings of the Sixth Annual
ACM International Conference on Mobile Computing and
Networking (MOBICOM), 2000; 32–43.
[16] N.J. Thomas, D.G.M. Cruickshank, D.I. Laurenson. Performance of a TDOA-AOA hybrid mobile location system. In
 2008 Curtin University of Technology and John Wiley & Sons, Ltd.
Asia-Pacific Journal of Chemical Engineering
Proceedings of IEE 3G Communication Technologies, 2001;
216–220.
[17] A. Ward, A. Jones, A. Hopper. IEEE Pers. Commun., 1997;
4(5), 42–47.
[18] B.H. Wellenhoff, H. Lichtenegger, J. Collins. Global Position
System:Theory and Practice, 4th edn, Springer Verlag: New
York, 1997.
Asia-Pac. J. Chem. Eng. 2008; 3: 589–596
DOI: 10.1002/apj
Документ
Категория
Без категории
Просмотров
1
Размер файла
175 Кб
Теги
experimentov, solutions, detected, wireless, stud, network, frequency, sensore, ranging, acoustic
1/--страниц
Пожаловаться на содержимое документа