Processing Top-k Monitoring Queries in Wireless Sensor Networks

Top-k monitoring queries are useful in many wireless sensor network applications. There is a well-known approach called FILA to process this kind of queries. Its basic idea is to install a filter at each sensor node to avoid unnecessary transmissions of sensor readings. FILA uses two algorithms to ensure the correctness and efficiency of the approach: a query reevaluation algorithm and a filter setting algorithm. In this paper, we propose improvements to each of these two algorithms. First, we propose a decentralized query reevaluation algorithm to reduce the communication cost of sending probe messages. Second, we propose a linear regression-based filter setting algorithm to improve the effectiveness of filters. Experimental results on real data traces show that our proposed improvements further enhance the performance of FILA in terms of network lifetime.