Correction of Nonlinear Frequency Sweep in Frequency-Modulated
Research Article Correction of Nonlinear Frequency Sweep in...
Transcript of Research Article Correction of Nonlinear Frequency Sweep in...
-
Hindawi Publishing CorporationInternational Journal of Distributed Sensor NetworksVolume 2013, Article ID 294967, 6 pageshttp://dx.doi.org/10.1155/2013/294967
Research ArticleCorrection of Nonlinear Frequency Sweep inFrequency-Modulated Continuous-Wave Laser Range Sensor
Soo-Yong Jung,1 Seong Ro Lee,2 and Chang-Soo Park1
1 School of Information and Communications, Gwangju Institute of Science and Technology, 261 Cheomdan-gwagiro, Buk-gu, Gwangju500-712, Republic of Korea
2Department of Information and Electronics Engineering, Mokpo National University, Yeongsan-ro, Cheonggye-myeon, Jeollanam-do,Muan-gun 1666, Republic of Korea
Correspondence should be addressed to Chang-Soo Park; [email protected]
Received 1 February 2013; Revised 10 June 2013; Accepted 11 June 2013
Academic Editor: Tai-hoon Kim
Copyright Β© 2013 Soo-Yong Jung et al.This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Wepropose a linearizationmethod for reducing the effect of nonlinear frequency sweep in a frequency-modulated continuous-wave(FMCW)based laser range sensor. In FMCWlaser range sensors, nonlinear frequency sweep can severely degrade themeasurementaccuracy because it gives the system ambiguity when determining the target range. In general, voltage controlled oscillators (VCO)which are used for frequency modulation show nonlinear frequency sweep property even though the input voltage signal is a linearramp signal. To solve this problem, we adopt an additional fixed delay structure to extract the nonlinearity and compensate it. Theproposed linearizationmethod has been worked out through the numerical process and the simulation, and this method effectivelyeliminates the nonlinear frequency sweep problem.
1. Introduction
Laser range sensors are remote distance sensing deviceswith typical applications, such as solid-target detections, 3Dvision, localization, and robotics [1β3]. To determine thedistance, projecting an optical signal onto an object andprocessing the reflected or scattered signal are performedin laser range sensors. Conventionally, pulsed time-of-flight(TOF), phase-shift measurement, and frequency-modulatedcontinuouswave (FMCW) are considered as major tech-niques for laser range sensors [4]. Pulsed TOF range sensordetermines the distance by measuring the round trip timeof the optical pulse signal. This method provides highsignal-to-noise ratio (SNR) and shows good performance;however high cost and large size are the drawbacks [5, 6].Phase-shift range sensor determines the distance using phasedifference between reference and reflected signals. Usingthis method, simple and low cost laser range sensor can berealized.However, intermediate frequency drift and influenceof the crosstalk degrade the performance of the phase-shifttechnique [7, 8].
To avoid these problems, we focused on the FMCW laserrange sensor. In an FMCW laser range sensor, a sinusoidalsignal with a constant rate of frequency change is transmit-ted, and this signal is reflected by a target. The frequencydifference between the transmitted and the reflected signals,called the beat frequency, contains the distance information.As long as the frequency sweep is linear, the beat frequencyis focused at a single frequency, and the target distance canbe easily extracted. However, in practice, the linear frequencysweep profile is not easy to obtain. If there is nonlinearityin the frequency sweep, the beat frequency is not focused ata single frequency, and it is difficult to determine the exactrange [9β11]. There are several techniques for linearizationof the frequency sweep, and most of them focused onthe linearization of the voltage controlled oscillator (VCO)frequency sweep because it is the crucial nonlinearity sourceof the system. The techniques are mainly of two types: one isopen-loop correction [12, 13] and the other one is closed-loopcorrection [14β16]. The open-loop correction method modi-fies the VCO tuning voltage properly to get a linear frequencysweep using a look-up table. However, since the frequency is
-
2 International Journal of Distributed Sensor Networks
controlled only by an open-loop system, inevitable frequencydrifts, due to temperature, environmental conditions, aging,and so forth, cannot be compensated for. The closed-loopcorrectionmethod adopts a phase locked loop (PLL) circuitryto linearize the frequency sweep. However, because PLLs aredynamic systems, any changes in the system input will causethe output not to follow immediately but to exhibit sometransient behavior. Such errors in the output frequency willhave a negative influence on the FMCWmeasurement result.
To solve these problems, we propose a method for cor-rection of frequency-sweep nonlinearity in a signal processorinstead of linearization of the VCO frequency sweep. Forlinearization, an additional fixed delay structure is adopted,and the frequency-sweep nonlinearity is extracted and usedfor compensation.We validate our linearizationmethod withnumerical analysis and computer simulation.
2. Nonlinearity Correction Method
In FMCW technique, a transmitter produces an FMCWsignal, that is, a sinusoidal signal with a constant rate offrequency change.This signal is backscattered by a target andgoes back to the emitter. Beat frequency which is frequencydifference between transmitted signal and backscattered sig-nal contains the distance information.
In intensity-modulated FMCW laser range sensor, alaser diode (LD) transmits an optical signal with directmodulation, and aVCO is adopted for frequencymodulation.The output signal frequency of theVCO is controlled by inputvoltage. If the input voltage ofVCO is a ramp signal, an opticalsignal with a constant rate of frequency increase is generated.This optical signal is backscattered by a target and goes backto a photo detector (PD) which can transfer optical signal toelectrical signal. Using a mixer and a low pass filter (LPF), asignal with beat frequency is extracted. Then the distance ofthe target can be expressed as
π =πRTT β π
2=πBEAT β π‘RAMP β π
2 β Ξπ, (1)
where πRTT is the round trip time of the optical signal, π is thespeed of light, πBEAT is the beat frequency, π‘RAMP is period ofthe ramp signal, andΞπ is frequency sweep range of theVCO.In third term of (1), π‘RAMP, π, and Ξπ are known parameters,and the beat frequency, πBEAT, is determined by the targetdistance. If the frequency sweep is linear, the beat frequency isconstant as long as the target range is not changed. However,in practice an exact ramp-like frequency profile is not easyto obtain. If there is nonlinearity in the frequency sweep, itmakes problem because the beat frequency is not constant,and it is difficult to find exact range.
To solve the nonlinear sweep problem, we propose equip-ping the FMCW laser range sensor with an additional fixeddelay structure, as depicted in Figure 1.
As shown in Figure 1, a fixed delay structure is addedfor linearization. When a voltage ramp signal is given to theVCO, the angular frequency of the VCO output signal can beexpressed as
π (π‘) = π0 + πΎ (π‘) β π‘, (2)
VCO
Delay
LPF
LPF
Frequencydetector
Frequencydetector
Signal processor
Fixed delay structure
PD
LDRamp generator Targ
et
LD: laser diode PD: photo detector
VCO: voltage controlled oscillatorLPF: low pass filter
Figure 1: Block diagram of the proposed FMCW laser range sensor.
where π0is the initial angular frequency and πΎ(π‘) means
the tuning rate to be compensated due to nonlinearity. Forsimplicity, assumingΞπΎ(π‘) βͺ π‘, the following equation showsthe phase of the VCO output signal:
π (π‘) = β«
π‘
0
π (π‘) ππ‘ = π0+ π0π‘ +
1
2πΎ (π‘) β π‘
2, (3)
where π0is the initial phase.Then, the VCOoutput signal can
be written as
πΈVCO (π‘) = πΎVCO cos (π (π‘)) , (4)
where πΎVCO is the amplitude of the VCO output signal.This signal is modulated to the laser diode and transmittedto the target. Practically, the LD output optical power is anonlinear function of the input electrical signal. Althoughwe can minimize the nonlinearity by modulating the inputsignal in the linear region of LD, the residual components canremain. The output from the LD is shown in (5) for a third-order nonlinearity, whereπ is the optical modulation index,andπ΄
2andπ΄
3are device dependent nonlinearity coefficients
[17]:
ππ(π‘) = π
0[1 + ππΎVCO cos (π (π‘))
+ π΄2π2πΎ2
VCOcos2(π (π‘))
+ π΄3π3πΎ3
VCOcos3(π (π‘))]
= π0[1 +
π΅2
2+ {π΅1+3π΅3
4} cos (π (π‘))
+π΅2cos (2π (π‘))
2+π΅3cos (3π (π‘))
4] ,
(5)
where π΅1
= ππΎVCO, π΅2 = π΄2π2πΎ2
VCO, and π΅3 =π΄3π3πΎ3
VCO. Then, the reflected signal detected by PD is
-
International Journal of Distributed Sensor Networks 3
expressed as [18]
πΈPD (π‘) = πΎπ ππ (π‘)
2
= πΎ
π π + 1
2ππ2ππ΄π sin2 (FOV) π
π(π‘ β πRTT)
2
= πΆ0+ πΆ1cos (π (π‘ β πRTT)) + πΆ2 cos (2π (π‘ β πRTT))
+ πΆ3cos (3π (π‘ β πRTT)) + πΆ4 cos (4π (π‘ β πRTT))
+ πΆ5cos (5π (π‘ β πRTT)) + πΆ6 cos (6π (π‘ β πRTT)) ,
(6)
where ππ (π‘) is the received optical signal, π is responsivity of
the PD,πΎ is constant of proportionality, π is a mode number,π is the target range, π is a reflection coefficient, π΄
π is the
photosensitive area of the PD, FOV is a field of view, andπRTT is the round trip time of the transmitted signal.TheVCOoutput signal and the reflected signal are multiplied througha mixer, and a signal with the beat frequency is obtained afterLPF, expressed as
πΈBEAT (π‘) =πΎVCOπΆ1
2cos (π (π‘) β π (π‘ β πRTT)) . (7)
A frequency detector is used for detecting the frequencyof the signal after LPF. A reciprocal frequency counter canbe used as the frequency detector. Because the reciprocalfrequency counter measures the period for one cycle ofthe waveform, it can support high resolution and very fastreadings. Also, a Schmitt trigger circuit can be used at inputstage of the frequency detector so that noise does not causespurious edges. Because frequency is derivative of phase, theangular frequency to be detected is expressed as
πBEAT (π‘) =π
ππ‘[π (π‘) β π (π‘ β πRTT)]
= πRTT β πΎ (π‘) + πRTTππΎ (π‘)
ππ‘π‘ β
πRTT2
2
ππΎ (π‘)
ππ‘.
(8)
As we can see in (8), if the tuning rate πΎ(π‘) is constant,then the angular frequency difference, πBEAT(π‘), is equal tothe product of the round trip time and the tuning rate.Because πBEAT(π‘) can be detected by the frequency detectorand the tuning rate is a known value, it is easy to find theround trip time, and we can extract the range of thetarget.However, usually, the frequency sweep of a VCO is notperfectly linear; that is, the tuning rate is not constant. Then,the beat frequency cannot give the exact round trip time.
To solve this problem, we added a fixed delay structurefor obtaining the tuning rate. The frequency detector can
detect the frequency difference between the VCO output anddelayed output signal, expressed as
πAUX (π‘) =π
ππ‘[π (π‘) β π (π‘ β π
π·)]
= ππ·β πΎ (π‘) + π
π·
ππΎ (π‘)
ππ‘π‘ β
π2
π·
2
ππΎ (π‘)
ππ‘,
(9)
where ππ·is the time delay in the fixed delay structure and
is much smaller than the frequency sweep time of the rampsignal. In (9),πAUX(π‘) can be detected by a frequency detectorwith the time delay. Thus, using the ordinary differentialequation, the tuning rate, πΎ(π‘), can be calculated as
πΎ (π‘) = exp(ββ« 22π‘ β π
π·
ππ‘)
β [β«2
2π‘ β ππ·
πAUX (π‘)
ππ·
exp(β« 22π‘ β π
π·
ππ‘) ππ‘]
= exp (β ln 2π‘ β ππ·)
β [β«2
2π‘ β ππ·
πAUX (π‘)
ππ·
exp (ln 2π‘ β ππ·) ππ‘]
=1
2π‘ β ππ·
β [2
ππ·
β«
2π‘ β ππ·
2π‘ β ππ·
πAUX (π‘) ππ‘] .
(10)
Because the time delay is much smaller than frequencysweep time of the ramp signal as mentioned above, π‘ is largerthan π
π·/2 in (10). Accordingly, the tuning rate can be
πΎ (π‘) =1
2π‘ β ππ·
β [2
ππ·
β«
π‘
ππ·/2
πAUX (π‘) ππ‘] . (11)
The derivative of the tuning rate can be obtained from (9)and expressed as
ππΎ (π‘)
ππ‘=
2
2π‘ β ππ·
(πAUX (π‘)
ππ·
β πΎ (π‘)) . (12)
Using (8) and (12), we can obtain a quadratic equation forπRTT, which is expressed as
1
2
2
2π‘ β ππ·
(πAUX (π‘)
ππ·
β πΎ (π‘)) π2
RTT
β {πΎ (π‘) +2π‘
2π‘ β ππ·
(πAUX (π‘)
ππ·
β πΎ (π‘))} πRTT
+ πBEAT (π‘) = 0.
(13)
The solutions of (13) are expressed as
πRTT =(2π‘/ (2π‘ β π
π·)) (πAUX (π‘) /ππ· β ππ·πΎ (π‘) /2π‘)
(2/ (2π‘ β ππ·)) (πAUX (π‘) /ππ· β πΎ (π‘))
Β±
β{(2π‘/ (2π‘ β ππ·)) (πAUX (π‘) /ππ· β ππ·πΎ (π‘) /2π‘)}
2β ((4 β πBEAT (π‘)) / (2π‘ β ππ·)) (πAUX (π‘) /ππ· β πΎ (π‘))
(2/ (2π‘ β ππ·)) (πAUX (π‘) /ππ· β πΎ (π‘))
.
(14)
-
4 International Journal of Distributed Sensor Networks
Frequency
Measurement time slot
Reference (transmitted)
signalReflected
signal
Ξf(=500
MH
z)
f0πRTT
fBEAT
t (ms)0.2 1
f0 + 500MHz
Figure 2: Simulation condition.
There are two solutions, and the solution in the detectablerange is selected. The obtained round trip time is constantover the measurement time, lying between the maximumround trip time and the frequency sweep time. In (14), π‘ canbe any value of the measurement time. The target range canbe obtained using a simple relation between distance and timeexpressed as
π =πRTT β π
2. (15)
3. Results and Discussion
We evaluate the proposed method using a computer sim-ulation. The frequency sweep range and sweep time ofa ramp signal are 500MHz and 1ms, respectively, and themeasurement time slot is 0.2 to 1ms. Figure 2 shows thesimulation condition.
To evaluate the proposed method, we modeled a non-linear frequency sweep and applied it to the FMCW laserrange sensor. Figure 3 shows three kinds of frequency sweeppatterns for each tuning rate.
We modeled one ideal linear and two kinds of nonlinearsweep patterns. Table 1 lists the tuning models.
When the target range is 30m, the beat frequencies andobtained target range for each case are shown in Figures 4(a)and 4(b), respectively. As shown in Figure 4(a), when thefrequency sweep is linear, the beat frequency is focused ata single frequency, 100 kHz, and the target range, 30m, canbe easily obtained. However, when the frequency sweep isnonlinear, it is difficult to extract the exact beat frequency,and the measurement accuracy is degraded. For case II,the obtained target range in the measurement time canvary between 15 and 34m, and, for case III, the obtained
0 0.2 0.4 0.6 0.8 10
1
2
3
4
5
Sweep time (s)
Freq
uenc
y (H
z)
Case I
Case II
Case III
Γ108
Γ10β3
Figure 3: Three kinds of frequency sweep patterns for each tuningrate.
Table 1: Tuning models.
Case Tuning rateI πΎ(π‘) = 500 Γ 109 (linear)II πΎ(π‘) = 500 Γ 107.5π‘β1/2
III πΎ(π‘) = 500 Γ 1012π‘
target range can vary between 12 and 58m, as depicted inFigure 4(b).
Two kinds of nonlinear frequency sweep patterns areapplied in the proposed method. Figures 5(a) and 5(b) showthe simulation results of the obtained target range plottedagainst the measurement time after linearization for case IIand case III, respectively. As we can see that, in both results,the target range is almost constant over the measurementtime and can be obtained successfully.
4. Conclusions
Wehave proposed a linearizationmethod using an additionalfixed delay structure in intensity-modulated FMCW laserrange sensor. For correction of the nonlinear frequency sweepproblem, a fixed structure was adopted to extract the tuningrate, and the target range was calculated using the obtainedtuning rate. We modeled three kinds of frequency sweeppatterns and applied them to the FMCW laser range sensor.When the frequency sweep was linear, the beat frequency wasfocused at a single frequency, and the target range was easilyextracted. On the other hand, when the frequency sweepwas nonlinear, there were multiple frequency componentsthat could not extract the exact target range. From theproposed linearization method, the simulation results clearlyshowed that the proposed method effectively eliminates thenonlinear frequency sweep problem and improves the meas-urement accuracy.
-
International Journal of Distributed Sensor Networks 5
0 1 2 3 4 50
0.5
1
1.5
2
Frequency (Hz)
Mag
nitu
de
Case III
Case I (linear)
Case II
Γ10β4
Γ105
(a)
0.0002 0.0004 0.0006 0.0008 0.0010
20
30
40
50
60
Case III
Case II
Case I (linear)Rang
e (m
)
Measurement time (s)
(b)
Figure 4: (a) Beat frequencies and (b) obtained target range for the tuning rate models.
20
15
25
30
30
30
30
30
30
Before linearization
After linearization
Rang
e (m
)
Rang
e (m
) 2 nm
0.0002 0.0004 0.0006 0.0008 0.0010Measurement time (s)0.0002 0.0004 0.0006 0.0008 0.0010
Measurement time (s)
(a)
20
30
40
50
60
30
30
30
30
30Before linearization
After linearization
Rang
e (m
)
Rang
e (m
) 2 nm
0.0002 0.0004 0.0006 0.0008 0.0010Measurement time (s)
0.0002 0.0004 0.0006 0.0008 0.0010Measurement time (s)
(b)
Figure 5: Simulation results after application of the proposed linearization method for the nonlinear frequency sweep of (a) πΎ(π‘) = 500 Γ107.5π‘β1/2 and (b) πΎ(π‘) = 500 Γ 1012 π‘.
-
6 International Journal of Distributed Sensor Networks
Acknowledgments
This work was supported by Priority Research Centers Pro-gram through the National Research Foundation of Korea(NRF), funded by the Ministry of Education, Science andTechnology (2009-0093828), and by the National ResearchFoundation of Korea (NRF) Grant funded by the Koreangovernment (MEST) (no. 2011-0017081).
References
[1] D. C. Carmer and L. M. Peterson, βLaser radar in robotics,βProceedings of the IEEE, vol. 84, no. 2, pp. 299β320, 1996.
[2] F. Blais, βReview of 20 years of range sensor development,βJournal of Electronic Imaging, vol. 13, no. 1, pp. 231β243, 2004.
[3] P. A. Forrester and K. F. Hulme, βLaser rangefinders,β Opticaland Quantum Electronics, vol. 13, no. 4, pp. 259β293, 1981.
[4] M.-C. Amann, T. Bosch, M. Lescure, R. MyllylaΜ, and M. Rioux,βLaser ranging: a critical review of usual techniques for distancemeasurement,β Optical Engineering, vol. 40, no. 1, pp. 10β19,2001.
[5] S. M. Nejad and S. Olyaee, βUnified pulsed laser range finderand velocimeter using ultra-Fast time-to-digital converter,βIranian Journal of Electrical & Electronic Engineering, vol. 5, no.2, pp. 112β121, 2009.
[6] P. PalojaΜrvi, K. MaΜaΜttaΜ, and J. Kostamovaara, βPulsed time-of-flight laser radar module with millimeter-level accuracy usingfull custom receiver and TDC ASICs,β IEEE Transactions onInstrumentation and Measurement, vol. 51, no. 5, pp. 1102β1108,2002.
[7] T. Bosch andM. Lescure, βCrosstalk analysis of 1 m to 10m laserphase-shift range finder,β IEEE Transactions on Instrumentationand Measurement, vol. 46, no. 6, pp. 1224β1228, 1997.
[8] G. Perchet and M. Lescure, βMagnification of phase shift fora laser rangefinder: intrinsic resolution and improved circuit,βJournal of Optics, vol. 29, no. 3, pp. 229β235, 1998.
[9] B. L. Stann, W. C. Ruff, and Z. G. Sztankay, βIntensity-modulated diode laser radar using frequency-modulation/continuous-wave ranging techniques,β OpticalEngineering, vol. 35, no. 11, pp. 3270β3278, 1996.
[10] T.-J. Ahn, J. Y. Lee, and D. Y. Kim, βSuppression of nonlinearfrequency sweep in an optical frequency-domain reflectometerby use of Hilbert transformation,βApplied Optics, vol. 44, no. 35,pp. 7630β7634, 2005.
[11] C. J. Karlsson and F. A. A. Olsson, βLinearization of thefrequency sweep of a frequency-modulated continuous-wavesemiconductor laser radar and the resulting ranging perfor-mance,β Applied Optics, vol. 38, no. 15, pp. 3376β3386, 1999.
[12] D. A. Williams, βA highly linearized mm-wave voltage con-trolled oscillator for FMCW radar applications,β IEE Collo-quium on Solid State Components for Radar, pp. 601β615, 1988.
[13] H.-G. Park, B. Kim, and Y. S. Kim, βVCO nonlinearity correc-tion scheme for a wideband FM-CW radar,β Microwave andOptical Technology Letters, vol. 25, no. 4, pp. 266β269, 2000.
[14] T. Musch, I. Rolfes, and B. Schiek, βA highly linear frequencyramp generator based on a fractional divider phase-locked-loop,β IEEE Transactions on Instrumentation and Measurement,vol. 48, no. 2, pp. 634β637, 1999.
[15] M. Pichler, A. Stelzer, P. Gulden, C. Seisenberger, and M.Vossiek, βFrequency-sweep linearization for FMCW sensors
with high measurement rate,β in Proceedings of the IEEEMTT-SInternational Microwave Symposium, pp. 1693β1696, June 2005.
[16] S. Scheiblhofer, S. Schuster, and A. Stelzer, βSignal modeland linearization for nonlinear chirps in FMCW radar SAW-ID tag request,β IEEE Transactions on Microwave Theory andTechniques, vol. 54, no. 4, pp. 1477β1483, 2006.
[17] X. N. Fernando and A. B. Sesay, βAdaptive asymmetric lin-earization of radio over fiber links for wireless access,β IEEETransactions on Vehicular Technology, vol. 51, no. 6, pp. 1576β1586, 2002.
[18] F. R. Gfeller and U. Bapst, βWireless in-house data communica-tion via diffuse infrared radiation,β Proceedings of the IEEE, vol.67, no. 11, pp. 1474β1486, 1979.
-
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Journal ofEngineeringVolume 2014
Submit your manuscripts athttp://www.hindawi.com
VLSI Design
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
DistributedSensor Networks
International Journal of