A Probabilistic Model Based-Tracking Method for mmWave Massive MIMO Channel Estimation

Accurate channel estimation with low pilot overhead is vital to exploit full benefit of Millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) system. Although compressive sensing (CS) based algorithms could solve this problem, their performances are affected by grid mismatch. To overcome these issues, a probabilistic model taking the spatial and temporal correlations into account for the massive MIMO is proposed in this correspondence. With this model, a quasi on-grid method is proposed to solve the problem of grid mismatch with low computational complexity. Finally, we integrate it into Turbo-CS framework and develop a message passing algorithm to solve the estimation problem. Simulation results reveal the superiority of proposed algorithm in estimation accuracy and tracking ability.