Exact Cramer-Rao Bounds for Semi-blind Channel Estimation in Amplify-and-Forward Two-Way Relay Networks

TL;DR

This paper derives exact Cramer-Rao bounds for semi-blind channel estimation in amplify-and-forward two-way relay networks, demonstrating significant accuracy improvements over training-based methods for certain modulation schemes.

Contribution

It provides the first derivation of exact CRBs for semi-blind estimation in this context, including simplified formulas for square QAM and high SNR approximations.

Findings

Semi-blind approach outperforms training-based estimation in accuracy.

Derived bounds applicable to PSK and square QAM modulation schemes.

Simplified expressions facilitate numerical implementation.

Abstract

In this paper, we derive for the first time the exact Cramer-Rao bounds (CRBs) on semi-blind channel estimation for amplify-and-forward two-way relay networks. The bounds cover a wide range of modulation schemes that satisfy a certain symmetry condition. In particular, the important classes of PSK and square QAM are covered. For the case square QAM, we also provide simplified expressions that lend themselves more easily to numerical implementation. The derived bounds are used to show that the semi-blind approach, which exploits both the transmitted pilots and the transmitted data symbols, can provide substantial improvements in estimation accuracy over the training-based approach which only uses pilot symbols to estimate the channel parameters. We also derive the more tractable modified CRB which accurately approximates the exact CRB at high SNR for low modulation orders.