無線感測網路之分散式偵測環境中基於對數相似比值之功率分配法則

Log-likelihood Ratio Based Power Allocation Scheme for Distributed Detection in Wireless Sensor Networks

林聖偉(Sheng-Wei Lin)

指導教授 ： 蔡育仁

國立清華大學/電機資訊學院/通訊工程研究所/碩士(2008年)

摘要

在本論文中，我們主要是在無線感測網路中處理分散式偵測環境的功率分配之研究。由於感測器本身受限於是用電池來供應運作所需的能量，再加上電池替換的困難，因此怎麼節省功率消耗一直是無線感測網路最重要的問題。所以我們提出一個新的功率分配方法就是每次感測器的傳輸功率是依據它當時所收到的觀察訊號強度。假如觀察訊號的對數相似比值之絕對值比較大，也就是觀察訊號所受到雜訊影響比較小，我們就給比較大的傳輸功率。相反的，假如觀察訊號的對數相似比值之絕對值比較小，也就是觀察訊號所受到雜訊影響比較大，我們就給比較小的傳輸功率。但是假如觀察訊號的對數相似比值之絕對值非常小時，也就是代表觀察訊號非常的不可靠，那麼我們就讓感測器不傳送訊息給中央處理器以便節省功率消耗。因此，我們考慮功率分配方法有『審查』要不要傳送訊號的概念。另外，我們也提出了當感測器傳輸功率受到限制的狀況。以及因為通道雜訊的方差不同，所以我們也考慮了當中央處理器會依據它所收到的訊號作不同的加權，期望感測器傳輸功率可以更低。最後，所提出的功率分配方法會跟均等傳輸功率方法比較，也就是所有感測器都傳送一樣的功率，且發現當同樣的偵測錯誤率限制下，我們提出的方法可以花費比較少的傳輸功率。

關鍵字

無線感測網路 ； 分散式偵測環境 ； 功率分配

並列摘要

In this thesis, we propose the power allocation schemes for distributed detection in wireless sensor networks and consider the amplifier gain based on the log-likelihood ratio (LLR) of the instantaneous observation signal received by the sensor node. If the absolute value of LLR is large, the instantaneous observation signal could be more reliable and the transmission power should be large. On the other hand, if the absolute value of LLR is too small, the instantaneous observation signal node could be less reliable and the transmission power should be small. But, if the absolute value of LLR is extremely small, we consider the sensor node should not transmit data to the fusion center so as to save power. Therefore, the idea of “censoring” is included in the proposed power allocation schemes. We also propose the power allocation scheme for the sensor nodes with the constrained transmission power since the transmission power of the senor node can not transmit with unlimited power. In addition, we consider the power allocation scheme with fusion weighting according to the different channel noise variances and expect the less total transmission power. Finally, the transmission power of the proposed schemes is compared to the equal power allocation scheme and find that the less total transmission power is required for achieving the same detection error probability.

並列關鍵字

sensor network ； distributed　detection ； power allocation

參考文獻

[1] I. Akyildiz, W. Su, Y. Sankarasubramniam, and E. Cayirci, “A survey on sensor networks,” IEEE Comm. Mag., vol. 40, pp. 102–114, Aug. 2002.

[2] H. Luo, J. Luo, Y. Liu, and S. K. Das, “Adaptive Data Fusion for Energy Efficient Routing in Wireless Sensor Networks”, IEEE Trans. on Computers., vol. 55, pp.1286 -1299, Oct. 2006.

[3] P. Sridhar, A. M. Asad, M. Jamshidi, “Hierarchical Data Aggregation in Spatially Correlated Distributed Sensor Networks”, World Automation Congress -International Symposium on Robotics and Applications, Budapest, July 2006.

[4] J. F. Chamberland and V. V. Veeravalli, “Wireless Sensors in Distributed Detection Applications,” IEEE Signal Process. Mag., vol. 24, pp. 16–25, May 2007.

[5] Z. Chair and P.K. Varshney, “Distributed Bayesian hypothesis testing with distributed data fusion,” IEEE Trans. Syst., Mag., Cybern., vol. 18, no. 5, pp. 695–699, Sept. 1988.