A Reduced Complexity Blind Estimation Method For Simo Channels Based On The Cross-Relations Subspace

In this paper, a power-based channel identification method is investigated based on the Cross-relations (CR) subspace of SIMO channels. The proposed method remains in stable condition and reduces the computational burden due to avoiding matrix inversion and joint square matrix multiplication. Compared to the Mul-tichannel least mean squared (MCLMS) method, it owns a faster convergence rate and the higher estimation accuracy with a close complexity. Furthermore, its performance is the same as the Cross-relations subspace iterations with Inversion (CRSI-INV) method and exceeds that of the Multichannel Newton (MCN) method in case of sufficient data samples, but with a lower complexity. Besides, a comparative study on channel identification methods is proposed and analyzed with respect to convergence rate and computational complexity. Simulations are also provided for demonstrating the superior performance of the proposed scheme.