Cluster-based time-domain channel prediction for dynamic wireless communications

System performance of mobile wireless communications is limited by fast channel variations, which are mostly due to the mobility of user terminals and scatterers. In order to obtain the optimal system performance, precise knowledge of the channel variations is often needed through channel prediction. A new cluster-based time-domain predictor with improved performance is proposed in this paper. Our predictor is based on the observation that multipath components within each cluster are highly correlated, while different multipath clusters share independent variations. The accuracy of the channel predictor is further enhanced by the proposed algorithm to separate the estimation noise and true channel variation information in each cluster. The principle and performance of the proposed channel predictor have been verified through numerical simulations.