Research On Interference Alignment In Ultra-dense Network

In this thesis, the interference characteristics of ultra-dense network are analyzed, and then the interference alignment method is applied to deal with the interferences in the ultra-dense scenario, and the acquisition problems of the channel state information are discussed in the practical application.Ultra-dense network is one of the key technologies of next generation mobile communication system on the deployment of network, which can greatly increase the system capacity per unit area. But the distance between the base stations is just 10 meters or smaller, which causes serious interference problems. Compared with traditional mobile communication network, the interference characteristics of the ultra-dense network are as follows. The interference intensity and quantity suffered at terminals rised dramatically and there’s a little difference in the intensity between the interference suffered at celledge terminals and central terminals. Therefore, the traditional interference suppression method is not suitable to solve the interference problem of ultra-dense network. So the interference alignment is used to restrain the interference in the ultra-dense network.The research background and several implementation methods of traditional interference alignment are introduced. The clustering interference alignment and partial interference alignment are introduced to reduce the complexity of the traditional iterative interference alignment greatly with low performance loss. The simulations of two algorithms in the ultra-dense network scenario show that the advantage of the sum rate performance is kept.Analog feedback and limited feedback are discussed to obtain the channel state information at the base stations in the practical application. Details of analog feedback and the limited feedback are introduced and the performance of two algorithms in the ultra-dense network are simulated. The simulation results show that two feedback algorithms can guarantee the system performance and reduce the feedback overload greatly.