Optimal power allocation for energy harvesting communications with limited channel feedback

Abstract

In energy harvesting (EH) wireless communications, full channel state information (CSI) at transmitter (TX) is generally needed to determine transmission power. In this paper, we propose a transmission power policy based on only 1-bit feedback of the CSI sent from the receiver (RX), for an EH communication system over Rayleigh fading channels. The RX periodically sends 1-bit feedback by comparing the channel power gain with a predetermined threshold. The TX adjusts transmission power based on the 1-bit feedback and its currently available energy. To obtain the optimal channel threshold and transmission power, we formulate a constrained optimization problem to maximize the throughput under an arbitrarily deterministic EH process. Constrained by the battery energy update process and channel state transition process, this problem follows a Markov decision process and can be solved via backward induction method. Simulation results show that our proposed policy has higher throughput performance than other policies.