Time-Varying Channel Estimation for Frequency-Division Duplex Massive MIMO Systems

The gain of massive multiple-input multiple-output (MIMO) systems relies on reliable channel estimation. Most of the traditional channel estimation schemes were designed for quasi-static situations which will suffer from significant performance degradation in scenarios with high user mobility. This paper investigated the time-varying channel estimation problem in frequency division duplex (FDD) massive MIMO systems. A soft-structured prior model based sparse Bayesian estimation (SSPM-SBE) scheme was proposed by exploiting the spatial-temporal property of the angular domain channel. The variational optimization framework was employed to solve the complex Bayesian estimation problem arisen in the proposed scheme. The SSPM-SBE can utilize both the current and previous received pilot signals to improve the estimation performance, and does not require the prior of large-scale channel information. The simulation results demonstrate the superiority of the proposed scheme.