| The fifth generation wireless communication system will meet the requirements of ultra-high peak rate, ultra-high connection density and ultra-high energy efficiency. Massive multiple input multiple output(MIMO) technology can make full use of space resources to enhance system performance, so it is recognized as one of the core technology of 5G in the future. However, the scene of multi-cell Massive MIMO will cause more serious interference and there is no good solution by far. In the frequency division duplexing(FDD) Massive MIMO system, the acquisition of channel state information(CSI) not only needs to occupy more resources, but also is still very complex. The joint processing of multi-cell can improve the system performance, but it will require a large number of inter-cell collaboration.In order to overcome the more complex interference problem in multi-cell FDD Massive MIMO system, this thesis use cloud radio access network(C-RAN) as the network framework and apply the improved joint spatial division and multiplexing(JSDM) technology in the system to reduce the complexity of precoding algorithm. This thesis also combines the antenna optimization to design a novel interference alignment algorithm, which will enhance the system capacity. The main works of this thesis are as follows:For the sake of reducing the overhead of CSI and effectively eliminating the interference of multi-cell MIMO system, the improved JSDM method which based on pseudo-inverse in multi-cell FDD systems is proposed. Firstly, according to their geographical geometry structure, the users in the cell are grouped so that all of them in the same group can have the same channel covariance matrix. Secondly, for the convenience of analysis, the multi-cell FDD Massive MIMO system model based on C-RAN architecture is transformed into a bidirectional interference channel model and a bidirectional multiple access channel model respectively. Thirdly, for the user group distributed in cell edge, the first-order precoding matrix based on pseudo-inverse method is designed to eliminate the inter-group interference caused by user group distributed in other cell edge. Lastly, the second-order precoding matrix is designed to eliminate the intra-group interference. The theoretical analysis and the simulation results show that the proposed scheme can not only reduce the algorithm complexity greatly and eliminate the interference effectively, but also improve the sum rate of system in some extent.Consider the problem that the complexity is too high and it needs a large number of antennas, when interference alignment is applied directly in multi-cell FDD Massive MIMO system, this thesis presents the interference alignment scheme based on JSDM. Firstly, in order to effectively reduce the number of transmitter antenna and receiver antenna on the basis of cancelling interference, a lower complexity antenna selection algorithm is designed in the base station. While reducing the number of RF chains, this algorithm will create a favorable condition for the combination of JSDM and interference alignment. Secondly, a joint receiver filter matrix is designed to overcome the problem that the number of antenna increasing fast with the number of users. Theoretical analysis shows that, while the number of users is large, the number of antennas required by the proposed scheme is significantly reduced, which provides a way of thinking for the practical application of interference alignment. The simulation results show that the simulation results are consistent with the theoretical analysis without affecting the performance of the interference alignment. |
PDF Full Text Request