Research On Iterative Interference Alignment Technology In Multi-User MIMO Wireless System

In recent years, with the rapid development of wireless communication technology, the interference management problems become a key issue and a restriction in the modem wireless communication system communications capacity and quality. Then Interference alignment technology arises as one of the effective means for solving the problem of spectrum resource scarcity and improving channel capacity. Therefore,this paper forces on goal function and gradient optimization of the iteration method of interference alignment algorithm for multi-user multi-input multi-output (MU-MIMO) system.At first, the thesis analyses three classic iterative methods of interference alignment algorithms including minimize interference leakage interference alignment algorithm,maximize Signal-to-Interference-plus-Noise-Ratio (SINR) interference alignment algorithm and joint signal and interference alignment algorithm. The condition of channel duality and the average rate of network are analyzed.Then on the basis of three above algorithms, the interference alignment iterative objective function optimization problem is taken into consider, analyses rank optimization problems of interference alignment. And for nuclear norm interference alignment algorithm ignores the desired signal power, matrix fraction interference alignment algorithm is proposed.The proposed algorithm takes both interference and signal matrices into account to ensure the maximum freedom and increases the network rate.At last, the thesis analyses interference alignment algorithms based on gradient optimization consider the general iterative interference alignment algorithm complexity, slow convergence speed of the objective function and a large number of redundant overhead caused by joint designing precoding matrix and receiver filter between transmitters and receivers, the method of gradient optimization study of interference alignment algorithm is analyzed.The interference subspace projection distance problem is analyzed, and study an algorithm via minimizing projector distances of interference subspaces which alleviates the overhead for designing precoding matrixes at one side. An improved interference alignment based on minimizing joint projector interference distances is proposed which achieves interference alignment by steepest descent algorithm based on complex Grassmann manifold. In the algorithm, it is not only minimizes the interference energy spill outside of the interfering subspace, but also reduce the signal power spilled out the signal subspace. In addition, the proposed method does not need the duality of channel and only optimize the precoding matrix at one side, reducing the overhead in the previous algorithms. Compared with the traditional interference alignment algorithms, the proposed algorithm can get a higher network rate and faster convergence rate than the algorithm via minimizing projector distances of interference subspaces.