Blind identification of the channel with common zeros

This paper firstly derives an ARMA model for channels with common zeros. Then, based on this model, the blind signal recovery problem is reduced to a quadratic optimization with integer constraints through the generalized complementary projection operator of the subspace spanned by a received data matrix. Finally, the recovered signals are used to identify the SIMO channels. The theoretical analysis and simulation results show that the generalized complementary projection algorithm proposed has the excellent performance in blind signal recovery and channel identification, whether the SIMO channels have common zeros or not, in or outside of the unit circle.