Robust knowledge-aided multipath channel identification based on partial filter information

The subspace method is an effective approach for blind channel identification. It works in the case when information such as the pulse shaping filter and the anti-aliasing filter responses are fully known. In practice, unknown perturbation may cause the transmitter/receiver filter response to be partially known, such as with I/Q imbalance and distortions of the filter due to environmental factors (temperature, humidity. etc.). Here we introduce two blind channel identification algorithms for two common situations, improving the performance of channel identification in cases when perturbation exists. Specifically, if the perturbation is totally unknown, we propose an iterative channel identification algorithm; for the situation in which some statistical knowledge of perturbation, such as the covariance of the perturbation, is known, we propose a robust knowledgeaided iterative channel identification algorithm to improve the estimation accuracy. Our simulation results demonstrate the performance of our new approaches.