Recursive Joint Channel Estimation and Signal Detection for Massive MIMO Systems

In this paper, a low-complexity recursive joint channel estimation and signal detection method is proposed for massive multiple-input multiple-output (MIMO) systems to improve both estimation and detection performance. Specifically, by approximating the autocorrelation matrix of the input data signal as a diagonal matrix, the complexity of the recursive least squares (RLS) channel estimation algorithm can be reduced. Then, based on RLS channel estimation, the complexity of signal detection can also be reduced by a two-stage Sherman formula. After that, the results of signal detection are then fed back to channel estimation and a weighting matrix is applyed for further performance improvement, which leads to the proposed weighted diagonalized recursive channel estimation and detection (WDRCED). Simulation results demonstrate that WDRCED improves the performance of both estimation and detection with reduced computational complexity.