Research On Compensation For Channel Non-reciprocity Based On Compressive Sensing

Time Division Duplex(TDD) and Frequency Division Duplex(FDD) are both available in mobile communication. Comparing with FDD system, one of the main advantages of TDD system is that it has channel reciprocity, which can effectively improve the spectrum efficiency.However, the channel reciprocity can be influenced by many reasons. Without compensation for channel reciprocity, the natural advantage will disappear in TDD systems. This thesis analyses the influences of channel estimation and time-variations from the points of system capacity and error. After that, based on the sparse feature of channel response, two compensation methods are proposed for TDD-MIMO system and massive MIMO system respectively. The main work of this thesis are as follow:(1) This thesis introduces the advantages and disadvantages of TDD system and FDD system briefly and highlights the advantage of channel reciprocity in TDD system. After that, some causes of channel non-reciprocity are discussed in this thesis, which include channel estimation error, radio frequency(RF) front-end imperfections, unsymmetrical interference pattern and time variation of channel. Many literatures about compensation for channel non-reciprocity take the influence of time variation into account while ignoring the deviation of channel estimation of UL directly. In this thesis, both two causes are took into consideration and compressive sensing is introduced into the compensation methods tactfully, which can effectively improve the system performance with low complexity.(2) This thesis focuses on the influence of time variation. Firstly, it quantifies the capacity of MIMO system with the help of Singular Value Decomposition(SVD). After that, this thesis compares the system capacity when the channel is time-invariant or time-variant. It can be found that the system capacity is damaged when the channel is time-variant. At last, this thesis proposes a compensation method for channel non-reciprocity based on compressive sensing in TDD-MIMO system. This method uses the sparse feature of channel response and improves the reconstruction algorithm in compressive sensing. Besides, it takes channel estimation error into account. Simulation results show that this compensation method can effectively improve the spectrum efficiency with low complexity.(3) This thesis introduces the technology of massive MIMO briefly and then analyzes the important role of channel reciprocity in large scale antenna system. After that, with the help of spatial-temporal correlation in massive MIMO system, a compensation method for channel non-reciprocity based on compressive sensing is proposed for large scale antenna system. This method also takes channel estimation error into account. As the simulation results show, this compensation method can effectively reduce the system error with low complexity.