Iterative sparse channel estimation and demodulation for blind equalisation

Abstract

The authors formulate a blind equalisation problem minimising the residual that is difference between the received signal and the predicted signal. Since both channel impulse responses and transmitted symbols are unknown, the authors regard the problem as a kind of dictionary learning problem. The authors divide the problem into two sub-problems and iteratively find the channel impulse response and the demodulated symbols. Since the wireless channels with a large bandwidth tend to have a sparse multipath structure, the authors apply the compressive sensing technique to the channel estimation sub-problem. Numerical results show that the proposed algorithm achieves the Genie-aided lower-bound performance.