The Research On Interference Alignment In Multi-user Wireless Communication Systems

With the development of wireless communication technology, multi-user MIMO system gradually replaced the single-user MIMO system, and then multiple access interference problems are been brought out. Some interference cancellation approaches can be summarized as follows: orthogonalize, treat as noise and decode. With the increase of users, spectrum efficiency will be declined. Interference alignment is a new technology which can make interference signals aligned to a finite dimensional space to maximize the transmission space of desired signals. With the increase of users, the capacity can increase linearly which makes it become an important research direction.Topic of this thesis is derived from the tutor s national natural science fund project. The paper first studies two kinds of alignment scheme, including interference vector alignment and lattice alignment. In 3 users, closed solution of interference alignment is given, and the algorithems are combined with zero-forcing detection or successive interference cancellation detection to eliminate interference. By simulating, the performance of lattice alignment is superior to the vector alignment, and successive interference cancellation detection is more effective than zero-forcing detection. Afterwards, iterative interference alignment algorithms are proposed, which based on different optimization criteria such as minimum leakage interference, maximum signal interference noise ratio and minimum mean square error. Actually, Most of the systems are unreasonable, that is to say, interference cannot be completely aligned. So precoding matrix and interference suppression matrix is iterated to align interference. By simulating, the minimum mean square error algorithm is the best in performance; maximum signal interference noise ratio algorithm is the second; minimum leakage interference algorithm is poor. Finally, this thesis studies closed solution of more than 3 users in reasonable system, it started from 4 users situation, and spread to multi-user. A new interference alignment scheme is proposed which can align part of interference, and left no interference signal space to desired signal, and eliminate the interference by ZF. That method need no iterative, and greatly reduces the computational complexity. By simulating, it can achieve the same performance with the traditional method in total rate and can get better BER performance in high SNR.