由於無線感測網路的感測節點易受到環境影響，造成錯誤資料在感測網路環境上傳播，如果這些異常資料未能即時地偵測出來，可能導致資料分析錯誤，由於無線環境的不穩定性及感測節點有限的記憶體，更增加異常資料偵測的挑戰性。在本論文中，我們針對於無線感測網路上分散式異常偵測技術進行研究。我們利用馬可夫隨意場(Markov Random Fields)來建構感測網路的資料模型；我們探討不同的相似函數來決定馬可夫隨意場中的臨近節點；我們使用可信度傳遞推論演算法(Belief Propagation)去偵測異常資料節點。模擬結果顯示，我們的方法，在配合適當的相似函數，能正確及有效地找出資料的異常節點。

Data cleaning is an important issue for wireless sensor networks, since wireless sensor networks operate in harsh and unreliable environments. In this research, we investigate an in-network cooperative data cleaning scheme for spatial correlated outliers in wireless sensor networks. High degree of spatial correlation in sensor data is used to infer missing value or correct value. First, we use Markov random field to model data dependency in wireless sensor networks, and use different similarity functions to design decision rules for choice of neighbors in the Markov random field. Then, belief propagation algorithm is proposed to compute and infer fault network data based on maximum a posteriori probability (MAP). To verify the performance of the proposed faulty data cleaning scheme, we run intensive computer simulations. The results of simulations show that the proposed data cleaning scheme can achieve high accuracy if the appropriate decision rule is chosen to determine neighbors in Markov random field model.

[1]S. Subramaniam, T. Palpanas, D. Papadopoulos, V. Kalogeraki and D. Gunopulos. “Online Outlier Detection in Sensor Data Using Non-Parametric Models” In VLDB, the 32nd international conference, Seoul, Korea, 2006
[2]Yongzhen Zhuang and Lei Chen. “In-network Outlier Cleaning for Data Collection Sensor Network” In the First International VLDB Workshop on Clean Databases (CleanDB), Seoul, Korea, 2006
[3]Xiang-Yan Xiao, Wen-Chih Peng, Chih-Chieh Hung and Wang-Chien Lee. “Using SensorRanks for In-Nnetwork Detection of Faulty Readings in Wireless Sensor Networks” In ACM, the 6th international workshop on Data engineering for wireless and mobile access, Beijing, China, 2007
[4]Bo Sheng, Qun Li, Weizhen Mao and Wen Jin. “Outlier Detection in Sensor Networks” In ACM, the 8th international symposium on Mobile ad hoc networking and computing, Montreal, Quebec, Canada, 2007
[5]Fang Chu, Yizhou Wang, D.Stott Parker and Carlo Zaniolo. “Data cleaning using belief propagation” In ACM, the 2nd international workshop on Information quality in information systems, NY, USA, 2005
[6]Fabrizio Angiulli and Fabio Fassetti. “Detecting Distance-Based Outliers in Streams of Data” In the 16th ACM conference on Conference on information and knowledge management, Lisbon, Portugal, 2007
[7]Jessica Lin, Eamonn Keogh, Pranav Patel and Stefano Lonardi. “Finding motifs in time series” In the 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining ,Edmonton, Alberta, Canada, 2002
[8]D. Janakiram, Adi Mallikarjuna Reddy V, and A V U Phani Kurmar. “Outlier Detection in Wireless Sensor Networks using Bayesian Belief Networks” In IEEE, First International Conference on Communication System Software and Middleware, 2006”
[9]Yongzhen Zhuang, Lei Chen, X. Sean Wang and Jie Lian. “A Weighted Moving Average-based Approach for Cleaning Sensor Data” In IEEE, the 27th International Conference on Distributed Computing Systems (ICDCS), 2007
[10]Mark D. Krasniewski, Padma Varadharajan, Bryan Rabeler, Saurabh Bagchi, and Y. Charlie Hu. Tibfit: “Trust index based fault tolerance for arbitrary data faults in sensor networks” In Proc. Of International Conference on Dependable Systems and Networks, pages 672-681, 2005
[11]Tony Sun, Ling-Jyh Chen, Chih-Chieh Han, and Mario Gerla. “Reliable sensor networks for planet exploration” In Proc. Of International Conference on Networking, Sensing and Control, pages 816-821, 2005
[12]Themistoklis Palpanas, Dimitris Papadopoulos, Vana Kalogeraki, and Dimitrios Gunopulos. “Distributed deviation detection in sensor networks” SIGMOD Rec., 32(4), 2003
[13]Bo Sheng, Qun Li, Weizhen Mao, and Wen Jin. “Outlier Detection in Sensor Networks” In the 8th ACM international symposium on Mobile ad hoc networking and computing, NY, USA, 2007
[14]Sharmila Subramaniam, Vana Kalogeraki and Themis Palpanas. “Distributed Real-Time Detection and Tracking of Homogeneous Regions In Sensor Networks” In IEEE, the 27th International Real-Time Systems Symposium (RTSS), 2006
[15]Chia-ming Chang “Outlier Detection for Spatial Correlated Data” MD. National Taiwan University of Science and Technology, 2007