A structured approach to optimization of energy harvesting wireless sensor networks

We analyze the data throughput maximization problem over fading channels for a energy harvesting wireless sensor network. The effective use of energy harvesting wireless sensor networks requires an apt understanding of the underlying environmental processes. Energy as a constrained resource must be carefully utilized so as to enable good performance through a horizon of epochs, trading off sensing performance and energy usage. We develop an algorithm for the allocation of power across sensors and time which is based on observations of the underlying structure of the optimization problem. We provide an analysis of particular features of the problem and present a maximum sum rate algorithm for generalized energy flow constraints, which proves to be optimal given known energy availability and fading coefficients. We present simulation results to validate the theory and algorithm.