Blind Channel Estimation and Symbol Detection for Multi-Cell Massive MIMO Systems by Expectation Propagation
Massive MIMO systems exploit the favorable propagation condition of the radio channel, whereby the vector-valued channels between the base station (BS) and the terminals become mutually orthogonal. This property is used in a recently-proposed channel estimation method for multi-cell massive MIMO systems based on the eigenvalue decomposition (EVD) of the correlation matrix of the received vectors. In this paper, we present a blind channel estimation and symbol detection scheme for multi-cell massive MIMO systems based on expectation propagation (EP). The proposed algorithm is initialized with the channel estimation result from the EVD-based method. It is shown that in our EP formulation, channel estimation and symbol detection are 'decoupled' in that EP iterations for channel estimation can be performed without the knowledge of the specific transmitted symbols. Therefore, channel estimation can be performed first followed by symbol detection. In particular, a liner symbol detection scheme such as zero-forcing (ZF) or minimum mean-squared error (MMSE) algorithm may be employed. Simulation results show that after a few iterations, the EP-based algorithm significantly improves the performance of the EVD-based method in both channel estimation and symbol error rate. Comparisons are also made with the results from a recently proposed blind detection scheme and it is shown that the proposed algorithm has better performance.
Publication Source (Journal or Book title)
IEEE Transactions on Wireless Communications
Ghavami, K., & Naraghi-Pour, M. (2018). Blind Channel Estimation and Symbol Detection for Multi-Cell Massive MIMO Systems by Expectation Propagation. IEEE Transactions on Wireless Communications, 17 (2), 943-954. https://doi.org/10.1109/TWC.2017.2772837