Distributed Threshold Selection for Aggregate Threshold Monitoring in Sensor Networks

Motivated by applications like sensor, peer to peer, ad hoc networks there has been growing interest in monitoring large scale distributed systems. In these applications typically we wish to monitor a global system condition which is defined as a function of local network elements parameters. In this paper, we study Aggregate Threshold Queries in sensor networks, which are used to detect when an aggregate value of all sensor measurements crosses a predetermined threshold. The major constraint in designing monitoring applications is reducing the amount of communication burden which is the dominant factor of energy drain in wireless sensor networks. In this study, we address the aggregate threshold monitoring problem by proposing a distributed algorithm to set local thresholds on each sensor node so that only those sensors whose measurements crosses their local thresholds commence communication. We adopt the FPTAS optimization formulation of the problem [1] and propose a distributed algorithm as the solution to the problem. Simulation results demonstrate the validity of the proposed distributed algorithm in attaining very close performance as the centralized scheme.