Analysis of the noise robustness problem and a new blind channel identification algorithm

Blind channel identification has generated much interest in signal processing and communications. Although existing cross relation based blind channel identification algorithm can achieve promising results, one of the drawbacks is the performance degradation in a noisy environment. In this work, we show that the degradation in convergence performance of MCLMS is due to an implicit constraint imposed by the cross relation cost function. This constraint requires the estimated impulse responses to be of the same energy which is often untrue in practice. We next propose a new algorithm exploiting revised cost function to improve the robustness of MCLMS to noise. Monte Carlo simulation results show that the proposed algorithm can gain significant improvement in steady-state performance.