On the precoder design of a wireless energy harvesting node in linear vector Gaussian channels with arbitrary input distribution

TL;DR

This paper investigates optimal precoder design for a wireless energy harvesting node in linear vector Gaussian channels with arbitrary input distributions, deriving new algorithms for offline and online power allocation to maximize mutual information.

Contribution

It introduces the Mercury Water-Flowing power allocation method and develops algorithms for both offline and online scenarios in energy harvesting communications.

Findings

Optimal precoder's left singular vectors align with channel eigenvectors.

Derived the Mercury Water-Flowing power allocation algorithm.

Proposed algorithms demonstrate effective mutual information maximization.

Abstract

A Wireless Energy Harvesting Node (WEHN) operating in linear vector Gaussian channels with arbitrarily distributed input symbols is considered in this paper. The precoding strategy that maximizes the mutual information along N independent channel accesses is studied under non-causal knowledge of the channel state and harvested energy (commonly known as offline approach). It is shown that, at each channel use, the left singular vectors of the precoder are equal to the eigenvectors of the Gram channel matrix. Additionally, an expression that relates the optimal singular values of the precoder with the energy harvesting profile through the Minimum Mean-Square Error (MMSE) matrix is obtained. Then, the specific situation in which the right singular vectors of the precoder are set to the identity matrix is considered. In this scenario, the optimal offline power allocation, named Mercury Water-Flowing, is derived and an intuitive graphical representation is presented. Two optimal offline algorithms to compute the Mercury Water- Flowing solution are proposed and an exhaustive study of their computational complexity is performed. Moreover, an online algorithm is designed, which only uses causal knowledge of the harvested energy and channel state. Finally, the achieved mutual information is evaluated through simulation.