Blind Turbo Channel Estimation of QAM Signals Exploiting Code Constraints

We present a blind turbo equalizer for QAM signals that exploits the statistics that a channel code imposes on the transmitted signal. The receiver is an adaptation of previous work presented in [1], where a BPSK signal is encoded by a binary code and blindly recovered in a turbo way. The main concept is to exploit the constraints imposed by the parity check equations of the code on the transmitted sequence for estimating the channel. Two algorithms are presented: the first one blindly estimates the channel without feedback of the channel decoder and the second one iteratively refines the channel estimation using the soft feedback of the channel decoder. The performance of the proposed system is encouraging. It shows that our blind channel estimator can reach the same BER as a Turbo equalizer with perfect channel knowledge of the channel impulse response.