Binary Energy Harvesting Channel with Finite Energy Storage

TL;DR

This paper analyzes the capacity of a binary energy harvesting communication channel with finite energy storage, modeling it as a state-dependent timing channel with causal transmitter information, and derives a single-letter capacity expression.

Contribution

It introduces a novel capacity characterization for an energy harvesting channel with finite battery, linking it to a timing channel with causal noise information.

Findings

Derived a single-letter capacity expression involving an auxiliary variable.

Evaluated achievable rates using noise concentration and lattice coding techniques.

Extended results to noiseless ternary channels.

Abstract

We consider the capacity of an energy harvesting communication channel with a finite-sized battery. As an abstraction of this problem, we consider a system where energy arrives at the encoder in multiples of a fixed quantity, and the physical layer is modeled accordingly as a finite discrete alphabet channel based on this fixed quantity. Further, for tractability, we consider the case of binary energy arrivals into a unit-capacity battery over a noiseless binary channel. Viewing the available energy as state, this is a state-dependent channel with causal state information available only at the transmitter. Further, the state is correlated over time and the channel inputs modify the future states. We show that this channel is equivalent to an additive geometric-noise timing channel with causal information of the noise available at the transmitter.We provide a single-letter capacity expression involving an auxiliary random variable, and evaluate this expression with certain auxiliary random variable selection, which resembles noise concentration and lattice-type coding in the timing channel. We evaluate the achievable rates by the proposed auxiliary selection and extend our results to noiseless ternary channels.