Pragmatic lifetime maximization of cooperative sensor networks via a decomposition approach

This paper addresses the problem of lifetime maximization under unequal and time-varying channel conditions, individual battery constraints, and estimation quality requirements at the fusion center. The standard tool for solving this problem (dynamic programming) has exponential complexity with number of sensors and states and needs heavy information exchange with sensors at each iteration. Also errors are introduced via coarse quantization of the parameters, which is forced by complexity concerns. In light of these issues, we propose a pragmatic method via a decomposition: The overall SNR requirement a is "divided" among sensors according to their battery powers and radio link statistics, and then individual sensors transmit powers are carefully controlled to maximize the lifetime. The proposed decomposition drastically reduces the computational requirement, and also allows a semi-distributed control of sensor transmit powers. Simulations verify the viability of this method.