A new method in blind estimation of fast time-varying channels based on subspace method

Traditional blind subspace-based channel estimation algorithms using second order statistics cannot estimate time-varying channels. However, there are some methods to modify these algorithms in order to estimate time-varying channels to some extent. Despite that, estimation error is high. We propose a method, based on ULV updating using modified Stewart sliding windows, to identify fast time-varying channels; it is able to estimate time-varying channels with low error and high speed convergence.