無線感測網路上利用記錄表之錯誤容忍即時決策融合機制

On-Line Fault-Tolerant Decision Fusion Scheme Based on a Record Table in Wireless Sensor Networks

吳政洋(Jeng-Yang Wu)

指導教授 ： 杜迪榕

共同指導教授 ： 王藏億

國立暨南國際大學/科技學院/資訊工程學系/碩士(2006年)

https://doi.org/10.6837/NCNU.2006.00012

摘要

無線感測網路是由大量的感測節點所組合而成的，其中的感測節點會在每個偵測時間點回報各自的local decision給一個fusion center，接著fusion center透過資訊融合的過程決定一個final decision，使得觀測者能了解監測環境的狀態。然而這些節點通常是價格便宜或被隨機部署在環境中，所以感測網路中往往會存在一些故障的感測節點。除此之外，故障的節點總是回報不可靠的資訊，當故障節點的影響力很大時，fusion center可能會作出錯誤的判斷。然而廣泛的故障節點型態，使得錯誤容忍偵測機制的設計相當困難，本研究提出故障節點隔離的機制，fusion center試圖透過連續性收到的local decision，偵測存在網路中的故障節點。本研究設計一個記錄表用來記錄感測節點回傳local decision的歷史，而fusion center利用記錄表裡的資訊，判定網路中行為特異的節點為故障節點。在decision fusion過程中，故障節點曾經回報的資訊將會被隔離，以改善分散式感測系統的偵測效能。因為無線感測網路的計算能力限制，本研究提出一種low-complexity的故障節點偵測機制，其使用簡單的統計方法迅速地偵測故障的節點。模擬的結果顯示本研究所提出的機制，其能有效地偵測網路中的故障節點。除此之外，當節點的故障型態顯著時，本研究所提出的機制對事件偵測之能力優於傳統的方法，傳統的方法指的是部署likelihood ratio資訊融合法則的分散式系統，在likelihood ratio運算的過程中沒有移除不可靠的local decision。

關鍵字：無線感測網路、錯誤容忍偵測、故障節點偵測、決策融合法則

關鍵字

無線感測網路 ； 錯誤容忍偵測 ； 故障節點偵測 ； 決策融合法則

並列摘要

Wireless sensor networks (WSN) are composed of a large number of sensor nodes. Each sensor node transmits individual local decision to a fusion center at every measured time step. The fusion then makes a final decision through the decision fusion process. Therefore, the observer can know the state in the monitored region of interest. However, the sensor nodes in WSN are very prone to damage due to the low-cost design and random deployment. Additionally, the faulty sensor nodes always report unreliable information. The fusion center may make a wrong decision when the combined effect of faulty nodes is great. Moreover, the wide range of the sensor fault types renders the design of fault-tolerant detection very difficult. This work addresses fault isolation in WSN where the fusion center attempts to identify faulty nodes through temporal sequences of received local decisions. A record table is designed to record the history of local decisions transmitted from all nodes in this study. The fusion then identifies a faulty node whose behavior is unusual through the information in the record table. The corresponding data transmitted by faulty nodes will be isolated in the decision fusion process in order to improve the detection performance of the distributed system. Because of the computing capability constraint in WSN, a low-complexity fault detection scheme, which adopts simple statistics to quickly identify faulty nodes, is proposed. The simulation results show the proposed approach is effective in identifying faulty members in a network. Additionally, when sensor faults significantly deviate from normal nodes, the hypothesis detection capability of the proposed scheme is superior to the conventional approach. The conventional approach refers to the distributed detection employing the likelihood ratio fusion rule without removing unreliable local decisions from the likelihood ratio computation.

Keywords: Wireless sensor networks, fault-tolerant detection, sensor fault detection, decision fusion

並列關鍵字

Wireless sensor networks ； fault-tolerant detection ； sensor fault detection ； decision fusion

參考文獻

[1]I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, “A survey on sensor networks,” IEEE Communications Magazine, vol. 40, no. 8, pp. 102–114, August 2002.

[2]S. A. Aldosari and J. M. F. Moura, “Detection in decentralized sensor networks,” in Proceedings of the 2004 IEEE International Conference on Accoustics, Speech, and Signal Processing, vol. 2, pp. 277–280, May 2004.

[3]M. Basseville, “Detecting changes in signals and systems-a survey,” Automatica, vol. 24, no. 3, pp. 309–326, May 1988.

[4]M. Basseville and I. Nikiforov, Detection of Abrupt Changes-Theory and Applications, Prentice-Hall, 1993.

[5]K. Chakrabarty, S. S. Iyengar, Hairong Qi, and Eungchun Cho “Grid coverage for surveillance and target location in distributed sensor networks,” IEEE Transactions on Computers, vol. 51, no. 12, pp. 1448–1453, December 2002.