Cluster-Based Fast Time-Varying MIMO Channel Fading Prediction in the High-Speed Scenario

Abstract

Accurate channel state information (CSI) is important for the coherent detection of multiple-input-multiple-output (MIMO) system. Especially in a high-speed scenario, fast time-varying CSI gotten by the conventional channel estimation schemes tend to be out of date, thus tracking and predicting CSI are more attractive and indispensable. Motivated by those, a time-varying MIMO channel fading prediction framework is proposed in this paper. The principle behind our scheme is that the cluster-based fading channel has the spatial consistency property, which means the small-scale parameters of channel clusters evolve continuously and smoothly in the time and spatial domains. Thus CSI can be tracked and predicted within several or tens of wavelengths. The proposed scheme is composed of an extended Bayesian Estimation Kalman Filter to track the time-varying CSI evolution, and a Cluster Drifting Based Prediction to obtain the small-scale parameters of channel clusters. The performance of the proposed scheme is simulatively verified by a standard clustered-based channel model.