Successive Interference Cancellation Based Channel Estimation for Massive MIMO Systems

This paper focuses on time shifted pilot (TSP) based time division duplex (TDD) massive multiple-input multiple-output (MIMO) systems. Since the performance of TSP with finite number of base station (BS) antennas is limited by poor channel estimation, the impact of channel estimation error is firstly analyzed, which demonstrates that the signal to interference plus noise ratios (SINR) of both downlink and uplink are inversely proportional to the covariance coefficient of the channel estimation error. Next, a successive interference cancellation based TSP (SIC-TSP) scheme is proposed where the downlink data contamination on the uplink pilot is cancelled out by utilizing the shared downlink data, precoding vectors and the estimated channels among BSs. To obtain accurate estimation for the channels among BSs, SIC-TSP employs a special time slot at the beginning of a frame, where each BS will transmit orthogonal pilots. Simulations are conducted to obtain the optimal uplink pilot power, and it is shown that the proposed SIC-TSP can reduce the normalized square channel estimation error by 11.3 dB, achieving a gain of 27% on both the downlink and uplink spectral efficiency compared with the existing TSP scheme.