Research On Channel Reciprocity Of TDD System And Its Application

Recently, with people’s growing demand on wireless communication, the next generation of mobile communication, LTE, is targeting on providing higher transmission rate and better quality of service. By utilizing MMO-OFDM technology, LTE system can significantly increase the spectrum efficiency and transmission rate. MIMO can adopt CSIT to precode data before transmission. This precoding technique can efficiently reduce interference and increase capacity of MIMO. In FDD mode, MSs transmit CQI to BSs through limited feedback. However, in TDD mode, since uplink and downlink transmission utilizes same frequency, they share the same propagation path, downlink CSI can be directly gained from uplink CSI. By utilizing channel reciprocity, TDD system can not only effectively reduce the overhead caused by feedback mechanism, but also can eliminate the quatization error.However, in real transmission scenario, many factors serves as constraints on channel reciprocity, which render it hard to implement. On one hand channel reciprocity exists only in coherence time span. On the other hand, Factors, like estimation error, time varying channel, imbalance of I/Q and imbalance interference between uplink and downlink, will render channel reciprocity useless. As a result, BS can hardly utilize reciprocity to obtain DL-CSI directly from UL-CSI. In this paper, from the perspective of MIMO capacity and bit error rate, the influence of time varying channel on reciprocity is thoroughly analyzed. Besides, this paper also puts putforward an adaptive prediction algorithm as a compensation scheme. It also considers the implementation of channel prediction with precoding scheme. The innovative achievements of this paper can be summarized as follows:Firstly, this paper gives a thorough quantitative analysis on how time varying channel can influence on channel reciprocity and puts forward the compensation scheme, which is based on channel prediction. By tracking the prediction error trend, the algorithm can adjust the prediction order and prediction step, which will not only largely reduce the algorithm complexity but also keep the prediction as accurate as possible.Secondly, in this paper, by considering TD-LTE frame structure, it puts forward a two-mode adaptive channel prediction scheme. When mobile station speed is low, frame-level prediction mode can be utilized; when the mobile station is under the condition of high speed, subframe-level prediction mode will be adaptively chosen. Moreover, by combining the proposed prediction scheme with non-codebook precodig techniques, the performance of MIMO system is thoroughly analyzed. In this paper, MMSE is chosen as codebook-free scheme to simulate the performance based on TD-LTE link level simulation platform.To conclude, there are two major contributions of this paper:this paper proposes adaptive channel prediction scheme to combat channel reciprocity loss, caused by time-varying channel. The proposed scheme can perfectly balance accuracy and complexity. Besides, based on TD-LTE frame structure, this paper also proposes an adaptive two-mode switching prediction scheme. Combining with MMSE codebook-free scheme, the proposed scheme efficiently increase system performance.