Blind Channel Estimation with Fewer Data Received

In this paper, a sophisticated blind channel estimation scheme is proposed for cyclic-prefix single-carrier systems. By rearranging the symbols within one received data vector, we expand the data vector with one column to a row-linearly- uncorrelated data matrix with two columns. Compared with the conventional subspace blind channel estimation, the method therefore can save half amount of the signals needed. This feature makes the proposed method be applicable for the cases where there exists limitation on the amount of the received data. Moreover, the advantage, which the complex scalar ambiguity existing in blind estimation could be reduced to the sign ambiguity, is inherited also. The whole process is dealt without the training sequences or redundant symbols interpolated. Simulations show the effects of the proposed method.