Semi-Blind Channel Estimation and Interference Alignment for Massive MIMO Network

In this paper, we analyze the performance of massive MIMO network with two-antenna users. The channel estimation is performed using the pilot sequence for one antenna and a simple data-aided estimation for the extra antenna. The advantage of the proposed method is that the channel estimation overhead does not increase linearly with the number of the user's antennas. We then compare between two schemes using this extra antenna: spatial repetition scheme and a simplified interference-alignment precoding scheme that does not require any coordination between the base stations or the users. It is shown that while the first scheme has the effect of virtually increasing (doubling) the number of antennas at the base station, the second precoding scheme outperforms the first one as it effectively decreases the size of the interference space. We analytically obtain a closed form for the achievable rate of one case using polynomial expansion approximations. The simulations verify the validity and the tightness of the obtained bound even at finite (not very large) number of BS antennas. The proposed algorithm is easily extended for more than two antennas.