Semi-blind channel estimation for amplify-and-forward two-way relay networks employing constant-modulus constellations

We present two channel estimation algorithms for amplify-and-forward two-way relay networks that employ constant-modulus constellations. The proposed algorithms do not assume complete knowledge of the transmitted symbols: they only require the transmission of a very short training sequence in order to resolve the inevitable phase ambiguity in the channel estimate. We first derive the maximum-likelihood (ML) estimator which is shown to perform very well even for a relatively small number of samples when the signal-to-noise ratio is sufficiently high. To address the high computational complexity of the ML estimator we propose a second, simpler, algorithm that can be updated at run-time and performs well for a sufficiently large sample size. Theoretical and experimental studies demonstrate the performance of the proposed algorithms.