Adaptive OFDM channel predictor based on multi-cluster model

A channel predictor based on multi-cluster model was presented for adaptive orthogonal frequency division multiplexing (OFDM) system. This predictor utilized the characteristic that several paths usually arrive in the same cluster in wireless propagation environment. The channel prediction process was divided into three stages including learning, estimation and prediction. Firstly the number and the delays of the clusters were estimated by sending plentiful pilots before communication. Then during the communication process, the number and the pattern of pilots can be adaptively adjusted. Combined with the auto-regress method, the future channel state information can be predicted. This predictor can reduce the number of pilots used and improve system spectral efficiency. This new predictor's performance has been proved by Monte-Carlo simulation results.