Channel estimation and performance of BICM-ID with signal space diversity over time-correlated Rayleigh fading channels

Signal space diversity (SSD) has been recently applied to improve the error performance of bit-interleaved coded modulation with iterative decoding (BICM-ID) systems over wireless fading channels. Previous works, however, only consider perfect channel estimation at the iterative receiver of the system. Performing channel estimation is an important task in a practical BICM-ID system and is the focus of this paper. To exploit the technique of SSD for improving channel estimation, a new method of using training sequences is presented. The proposed method inserts pilot bits between the coded bits, i.e., before constellation mapping and signal rotation. It is therefore can be considered as superimposed training. Developed is a soft iterative channel estimator to work with the new arrangement of the training signal. The performance of the system with the proposed training and channel estimation is shown to be superior that with pilot symbol assisted modulation (PSAM). To benchmark the achievable error performance of the system with channel estimation, an analytical bound of the asymptotic bit error probability for BICM-ID using SSD over correlated fading channels and perfect channel state information (CSI) is also derived.