Modulation optimization for energy harvesting transmitters with compound Poisson energy arrivals

With the development of the energy harvesting technology, communication devices nowadays can be powered by the electrical energy obtained by converting different forms of energy from their ambience. The energy that becomes available to such transceivers varies both in time and amount, the exact information of which is usually unknown to the transceivers. We consider in this work the point-to-point communication between an energy harvesting transmitter and a receiver over a block-fading channel, where the transmitter has statistical and causal knowledge about the energy arrivals as well as the channel conditions. The stochastic energy arriving process is assumed compound Poisson, which provides both good mathematical tractability and enough physical generality. With the objective of maximizing the average throughput over a long operation time, we model the system as a Markov decision process and apply the policy-iteration algorithm to optimize the transmission policies with respect to all discretized system states. Several transmission strategies are proposed and compared from the aspects of ease of control, performance, and computational complexity.