A subspace projection method for blind identification using shifted correlation matrices (SPAWC-2004)

We present a subspace projection method for blind MIMO FIR channel identification by using the shifted correlation matrices. It is shown that the shifted second order correlation matrices can render us the partial column space of the channel convolution matrix. Subsequently, by employing an orthogonal projector constructed by the estimated partial column space of the channel convolution matrix, we developed an algorithm for blind system identification. Computer simulation results are included to validate the performance of the proposed algorithm.