TY - GEN
T1 - Distributed adaptive quantization for wireless sensor networks
AU - Fang, Jun
AU - Li, Hongbin
PY - 2007
Y1 - 2007
N2 - We investigate the problem of distributed parameter estimation under the most stringent bandwidth constraint that each sensor quantizes its local observation into one bit of information. Conventional fixed quantization (FQ) approaches, which employ a fixed threshold for all sensors, incur an estimation error growing exponentially with the difference between the threshold and the unknown parameter to be estimated. To address this difficulty, we propose a distributed adaptive quantization (AQ) approach, where, with sensors sequentially broadcasting their quantized data, each sensor adaptively adjusts its quantization threshold using prior transmissions from other sensors. Specifically, three adaptive schemes are presented in this paper. The maximum likelihood (ML) estimators associated with these three AQ schemes are developed and their corresponding Cramér-Rao bounds (CRBs) are analyzed. The analysis shows that our proposed one-bit AQ approach can asymptotically attain an estimation variance as least as only π/2 times that of the clairvoyant sample-mean estimator using unquantized observations. Numerical results are illustrated to show the effectiveness of the proposed approach and to corroborate our claim.
AB - We investigate the problem of distributed parameter estimation under the most stringent bandwidth constraint that each sensor quantizes its local observation into one bit of information. Conventional fixed quantization (FQ) approaches, which employ a fixed threshold for all sensors, incur an estimation error growing exponentially with the difference between the threshold and the unknown parameter to be estimated. To address this difficulty, we propose a distributed adaptive quantization (AQ) approach, where, with sensors sequentially broadcasting their quantized data, each sensor adaptively adjusts its quantization threshold using prior transmissions from other sensors. Specifically, three adaptive schemes are presented in this paper. The maximum likelihood (ML) estimators associated with these three AQ schemes are developed and their corresponding Cramér-Rao bounds (CRBs) are analyzed. The analysis shows that our proposed one-bit AQ approach can asymptotically attain an estimation variance as least as only π/2 times that of the clairvoyant sample-mean estimator using unquantized observations. Numerical results are illustrated to show the effectiveness of the proposed approach and to corroborate our claim.
KW - Adaptive quantization (AQ)
KW - Distributed estimation
KW - Wireless sensor networks (WSNs).
UR - http://www.scopus.com/inward/record.url?scp=50249095083&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=50249095083&partnerID=8YFLogxK
U2 - 10.1109/ACSSC.2007.4487452
DO - 10.1109/ACSSC.2007.4487452
M3 - Conference contribution
AN - SCOPUS:50249095083
SN - 9781424421107
T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers
SP - 1372
EP - 1376
BT - Conference Record of the 41st Asilomar Conference on Signals, Systems and Computers, ACSSC
T2 - 41st Asilomar Conference on Signals, Systems and Computers, ACSSC
Y2 - 4 November 2007 through 7 November 2007
ER -