Energy Efficient Top-k Query Processing in Dynamic Sensor Network

Sensor networks may generate a large amount of data during the monitoring process. It is crucial to conserve energy when processing these data. The detected sensory data vary in long monitoring period. The problem is how to design query processing with minimal energy and obtain correct result. In this paper, we propose a history-based approach to optimizing top-k query processing in sensor network and design threshold-estimate-prune-query algorithm. Energy consumption can be reduced by pruning unnecessary sub-queries and query message can be guided to right direction by cached data. Subset of the sensor network will respond the query. We design local-expand-query method to get more correct data when the environment changed. Simulation results show that the number of queried nodes can be significantly reduced in top-k query processing. The expand query can get more accuracy data when the environment varies and data distribution changed.