J-Divergence Based Decentralized Power Allocation Scheme for Distributed Detection in Wireless Sensor Networks

Abstract

AbstractIn a wireless sensor network with radar-like sensors, the allocation of the limited power of a sensor node between sensing module and communication module should be deliberated. Obviously, the energy consumption of the whole system can be lowered by jointly optimizing the detection performance of the sensor node and the communication performance between the sensor nodes and the fusion center. Therefore, the J-divergence between the distributions of the received local decisions at the fusion center under different hypotheses is used as a performance criterion to optimize the power allocation of sensor node. Numerical simulations shows that the proposed method can strike a good tradeoff between the communication channel quality and the detection quality of local sensor. Although the proposed method is suboptimum, it outperforms the best allocation scheme obtained by the uniform power allocation scheme and the identical symbol error-rate based Power Allocation under the typical fusion rules considered. Besides, the proposed method is independent of specific fusion rule and only needs local processing, which are another two key characteristics.Keywordsdistributed detectionpower allocationsignal detectionwireless sensor networks