Weighted compressive sensing based uplink channel estimation for time division duplex massive multi-input multi-output systems

摘要

In this study, the channel estimation problem for the uplink massive multi-input multi-output (MIMO) system is considered. Motivated by the observations that the channels in massive MIMO systems may exhibit sparsity and the channel support changes slowly over time, the authors propose one efficient channel estimation method under the framework of compressive sensing (CS). By exploiting the channel impulse response (CIR) estimated from the previous orthogonal frequency division multiplexing symbol, they firstly estimate the probabilities that the elements in the current CIR are non-zero. Then, they propose the probability-weighted subspace pursuit algorithm exploiting these probability information to efficiently reconstruct the uplink massive MIMO channel. Moreover, noting that the massive MIMO systems also share a common support within one channel matrix due to the shared local scatterers in the physical propagation environment, an antenna collaborating method is exploited for the proposed method to further enhance the channel estimation performance. Simulation results show that compared to the existing CS methods, the proposed methods could achieve higher spectral efficiency as well as more reliable performance over time-varying channel.