Enhancing Energy Efficiency for Routing in Sensor Networks with a Detection Mission

We consider the problem of energy-efficient routing for optimal detection in sensor networks. The spatial correlation structure of sensor measurements is modeled by a Markov random field (MRF), where the neighborhood system is defined by means of a nearest-neighbor dependency graph. The routing scheme that achieves the best detection accuracy and minimizes the total energy consumption is desired. Two techniques that enhance performance on energy efficiency of an existing routing scheme (DFMRF) are proposed. The techniques are motivated by exploring wireless multi-receiver diversity and redundancy in full coverage of the spanning tree respectively. Both theoretical analysis and simulation results are provided to verify the superiority of the proposed techniques over the original DFMRF in terms of energy efficiency. It is also shown that applying both techniques together is not always the optimal solution.