Information-Theoretic Analysis of an Energy Harvesting Communication System

TL;DR

This paper analyzes the capacity of energy harvesting communication systems with stochastic energy arrivals, showing it equals the capacity under an average power constraint, and proposes optimal power allocation strategies.

Contribution

It establishes the capacity equivalence for AWGN channels with stochastic energy arrivals and introduces two capacity-achieving schemes, along with optimal offline power allocation methods.

Findings

Capacity equals that of an AWGN channel with average power constraint.

Proposed save-and-transmit and best-effort-transmit schemes achieve capacity.

Derived optimal offline power allocation for time-varying energy arrivals.

Abstract

In energy harvesting communication systems, an exogenous recharge process supplies energy for the data transmission and arriving energy can be buffered in a battery before consumption. Transmission is interrupted if there is not sufficient energy. We address communication with such random energy arrivals in an information-theoretic setting. Based on the classical additive white Gaussian noise (AWGN) channel model, we study the coding problem with random energy arrivals at the transmitter. We show that the capacity of the AWGN channel with stochastic energy arrivals is equal to the capacity with an average power constraint equal to the average recharge rate. We provide two different capacity achieving schemes: {\it save-and-transmit} and {\it best-effort-transmit}. Next, we consider the case where energy arrivals have time-varying average in a larger time scale. We derive the optimal offline power allocation for maximum average throughput and provide an algorithm that finds the optimal power allocation.