Robust Transceiver with Tomlinson-Harashima Precoding for Amplify-and-Forward MIMO Relaying Systems

TL;DR

This paper proposes a robust transceiver design for multi-hop AF MIMO relaying systems using Tomlinson-Harashima precoding, effectively mitigating channel estimation errors and improving performance.

Contribution

It introduces a joint Bayesian robust design framework for THP, linear forwarding, and equalization, utilizing multiplicative convexity and matrix-monotone functions for optimal structure derivation.

Findings

Robust transceiver outperforms non-robust designs in simulations.

Optimal structure reduces design complexity to scalar variables.

Effective mitigation of channel estimation errors in MIMO relaying systems.

Abstract

In this paper, robust transceiver design with Tomlinson-Harashima precoding (THP) for multi-hop amplify-and-forward (AF) multiple-input multiple-output (MIMO) relaying systems is investigated. At source node, THP is adopted to mitigate the spatial intersymbol interference. However, due to its nonlinear nature, THP is very sensitive to channel estimation errors. In order to reduce the effects of channel estimation errors, a joint Bayesian robust design of THP at source, linear forwarding matrices at relays and linear equalizer at destination is proposed. With novel applications of elegant characteristics of multiplicative convexity and matrix-monotone functions, the optimal structure of the nonlinear transceiver is first derived. Based on the derived structure, the transceiver design problem reduces to a much simpler one with only scalar variables which can be efficiently solved. Finally, the performance advantage of the proposed robust design over non-robust design is demonstrated by simulation results.