A sliding window approach for dynamic event-region detection in sensor networks

Event region detection is an important problem in wireless sensor networks. However, a fundamental assumption required for the event-region detection schemes proposed in the literature is that the event-region under detection is quasi-static, i.e., the event-regions do not change during the detection period. This assumption could fail in detecting fast varying events, such as forest fire events and mudslides. Therefore, the detection task considered in the present study is to cope with the continuously varying event regions. Specifically, we adopt a sliding window approach, in which only the sensor measurements from the present to a fixed length time ago are utilized for making decision at any particular time slot. Our simulation results demonstrate the advantages of the proposed detection scheme.

[1]  Venkatesh Saligrama,et al.  Detection and Localization in Sensor Networks Using Distributed FDR , 2006, 2006 40th Annual Conference on Information Sciences and Systems.

[2]  Anil K. Jain,et al.  A Markov random field model for classification of multisource satellite imagery , 1996, IEEE Trans. Geosci. Remote. Sens..

[3]  Venkatesh Saligrama,et al.  Adaptive statistical sampling methods for decentralized estimation and detection of localized phenomena , 2005, IPSN 2005. Fourth International Symposium on Information Processing in Sensor Networks, 2005..

[4]  Aleksandar Dogandzic,et al.  Distributed Estimation and Detection for Sensor Networks Using Hidden Markov Random Field Models , 2006, IEEE Transactions on Signal Processing.

[5]  Michèle Basseville,et al.  Detection of Abrupt Changes: Theory and Applications. , 1995 .

[6]  Stan Z. Li,et al.  Markov Random Field Modeling in Image Analysis , 2001, Computer Science Workbench.

[7]  S. Sitharama Iyengar,et al.  Distributed Bayesian algorithms for fault-tolerant event region detection in wireless sensor networks , 2004, IEEE Transactions on Computers.

[8]  Qi Cheng,et al.  Collaborative Event-Region and Boundary-Region Detections in Wireless Sensor Networks , 2008, IEEE Transactions on Signal Processing.