Data-assisted channel estimation for uplink massive MIMO systems

A novel data-assisted channel estimation scheme is proposed for uplink massive multiple-input multiple-output (MIMO) systems to alleviate the performance bottleneck due to pilot contamination. Specifically, the uplink resource within a fading block is first divided into a number of data blocks, which are encoded respectively; then the detected data blocks are utilized iteratively to refine the channel estimation, improving the signal-to-interference-plus-noise ratio (SINR) of the following data blocks. In the existing literature, the system performance can not scale up with the number of base station´s antenna due to the effect of pilot contamination. The proposed scheme provides a promising approach to improve the uplink SINR by the order θ(LM/L+M) without increasing the length of pilot sequence, where L and M are the number of transmission symbols within a fading block and the number of base station´s antennas respectively. It is shown by simulations that the uplink performance of massive MIMO systems is significantly increased by the proposed scheme.