MIMO-OFDM downlink channel prediction for TDD systems using kalman filter

Abstract

Multiple-input multiple-output (MIMO) technique combined with orthogonal frequency division multiplexing (OFDM) has been considered as a key technology for next-generation wireless communication systems due to its potential of achieving high data rates and mitigating the hostile channel selectivity. In particular, MIMO downlink systems, such as beamforming or precoding system provide higher link capacity and simplify the receiver with channel state information (CSI) at the transmitter. In order to obtain CSI at the transmitter, a feedback from the receiver can be used when the up- and downlink channels are uncorrelated. In time-division duplex(TDD) systems, CSI for downlink can be obtained from uplink channel using reciprocity. However, the CSI from uplink is not accurate enough to keep track of the continuously varying channel characteristic in downlink period. A few channel prediction schemes, such as linear prediction and second order extrapolation are proposed to overcome the time difference between up- and downlink for TDD systems. Generally, second order extrapolation outperforms linear prediction in noiseless environment. However, second order extrapolation might have even worse performance than linear prediction because of noisy measurement data. Thus, a channel predictor which can resolve these problems due to outdated and imperfect CSI, is necessary. In this thesis, a MIMO-OFDM downlink channel prediction technique based on Kalman filter is proposed for TDD systems. The proposed method consists of three procedures: MMSE channel estimation, Kalman filtering and prediction, and linear interpolation. Kalman filter is employed to filter the estimated channel and to predict the next channel sample to determine the precoding weights. Simulation results of the proposed scheme based on IEEE802.16e system demonstrate that the proposed method improves the bit error rate (BER) performance significantly.