Maximizing the lifetime of energy harvesting wireless sensor networks

Lifetime is considered as one of the most important issues in designing wireless sensor networks (WSNs) because of the short battery lifetime, and the autonomous operation for a long period of time. Therefore, there has been much research interest in maximizing the lifetime of the WSN in the last few years. Renewable energy provides a promising way to extend the lifetime of WSN significantly, but one has to deal with the problem of highly dynamic renewable energy supply and sensor energy consumption. In this dissertation we address this problem using three approaches.;This dissertation first introduces a linear programming formulation for the problem of maximizing alpha-lifetime of the WSN using renewable energy sources. The alpha-lifetime of a sensor network is defined as the time duration in which alpha-percentage of sensory data is collected by the base station. The proposed formulation considers varying distributions of solar energy supply rate at different sensor nodes and jointly optimizes the transmission power of the sensor nodes and data routing to maximize alpha-lifetime. We investigate the relationship between the alpha-lifetime with the distribution of the solar energy supply rate, and the alpha-percentage values. This investigation is significant in practical deployment of the WSNs.;The dissertation next investigates how to maximize lifetime sequences of the WSN and the impact of renewable energy sources on the lifetime. The lifetime sequence is defined as a sorted list of lifetimes of sensors in non-decreasing order. We provide the first optimization formulation for maximizing both unconditional and conditional lifetime sequences. Our investigation shows that the lifetime of sensors that die later can be extended by renewable energy much more significantly than those that die early.;Finally, this dissertation proposes an optimization formulation scheme to maximize the WSN lifetime called Multi-Stage Optimization. Our proposed scheme considers the problem of multi-stage optimization as set of optimizations of single stages. Each single stage consists of two phases of optimization: Phase-1 optimizes the given problem of maximizing the lifetime, and Phase-2 re-optimizes the problem at a particular fraction of the optimized lifetime corresponding to the current stage. We show that our scheme is more efficient and more realistic than the previous works where they deal with the problem of optimization as single stage optimization.