Non-Myopic Energy Allocation for Target Tracking in Energy Harvesting UWSNs

For target tracking in underwater wireless sensor networks (UWSNs), how to improve energy utilization efficiency and target tracking accuracy with limited energy constraints is an important problem. Recent years, energy harvesting device has been developed and applied to UWSNs to guarantee energy supply, and reasonable and efficient energy allocation schemed are important. In this paper, energy allocation problem for target tracking in UWSNs is studied. The goal in this paper is to improve tracking accuracy under limited energy harvesting constraints. First, to maximize the overall accuracy in the whole process, the accumulated fisher information matrix is derived in a non-myopic way and used as performance metric. Second, based on the fact that energy consumption and tracking accuracy mainly depend on bit number of quantized measurement, an optimization problem is proposed to solve tradeoff between energy allocation and tracking accuracy under energy harvesting constraints. Third, to obtain optimal energy allocation for each time, the problem is formulated as a Markov decision process, which is solved by a dynamic programming algorithm in pseudo-polynomial time. The simulation results are presented to verify the effectiveness of our proposed scheme.