FP-CPNNQ: A filter-based protocol for continuous probabilistic nearest neighbor query

Abstract

An increasing number of applications in environmental monitoring and location-based services make use of large-scale distributed sensing provided by wireless sensor networks. In such applications, a large number of sensor devices are deployed to collect useful information such as temperature readings and vehicle positions. However, these distributed sensors usually have limited computational and communication power and thus the amount of sensor queries should be reduced to conserve system resources. At the same time, data captured by such sensors is inherently imprecise due to sensor limitations. We propose an efficient probabilistic filter-based protocol for answering continuous nearest neighbor queries over uncertain sensor data. Experimental evaluation on real-world temperature sensing data and synthetic location data showed a significant reduction in the number of update messages.