Algorithms for Energy-Efficient Query-Reduction in Wireless Sensor Networks

A wide area sensor system is a complex, dynamic, resource constrained collection of sensing devices. Typically, queries are sent out by master node(s) and query-satisfying data are aggregated at optimized points in a sensor network and sent back to master node(s). If patterns of queries sent out by master node(s) are studied, replies can be sent back based on patterns observed. This would be especially beneficial in saving energy for large heterogeneous sensor network whose nodes' location is partially unknown and could change. The contribution of this paper is to propose two simple learning algorithms (block based, incremental learning) to learn about the pattern of queries and also to use an associated sensor and data (ASD). Initial results show promising energy savings with this scheme when compared to flooding.