Interference Alignment Algorithms In Wireless Networks

With the continuing and rapid development of wireless communications, the scarcity of spectrum has resulted in a huge development of a vast variety of techniques to improve spectrum utilization, including MIMO (Multiple-Input Multiple-Output) systems, which have attracted more and more attention. In addition to interference among antennas, multiuser multi-cell MIMO systems will also bring interferences among bases stations as well as users. As a new viable tool to deal with interference, interference alignment (IA) has changed the conventional view that wireless networks based on MIMO are interference-limited systems, so it has been regarded as one of the hot research topics. The channel state information of transmitters (CSIT) is a very important factor to affect the performance of IA. In this paper, the following aspects have been investigated:1) A heterogeneous block fading channel model has been used to study the IA when there is no CSIT, including blind IA and staggered antenna switching IA. The first scheme is applied to the X network and the interference channel. The last one is applied to the MISO broadcast channel and X network. The simulation result shows that, degree of freedom (DoF)2M/(M+1) is achievable in the M×2X channel, K/2in the K-user interference channel, and MK/(M+K-1) in the MISO BC(Multiple-Input Single-Output Broadcast Channel) where the transmitter is equipped with M antennas and there are K receivers with each one equipped with single antenna.2) There often existes estimation error of the channel state information at transmitter in actual wireless systems, in order to analyze the impact of the error on the IA performance, five IA algorithms have been investigated based on Gauss-Markov channel error model, including alternating minimization algorithm (Alt-Min), asymptotic interference alignment (Asy-IA), maximum signal to interference and noise ratio (Max-SINR), minimum weighted leakage interference (Min-WLI) and IA based on minimum mean-square error criterion. Simulation shows that, channel imperfection has practically no effect on the sum rate for β≤0.01.3) In fast time-varying channel, the current channel state is completely independent with the past state, but the transmitter could improve the performance of the system by using delayed CSI which is gottent through feedback. How to use these delayed information to harvest the performance gains of the MIMO BC, X network and interference channels? The problem has been solved in this part. The simulation results show that, the DoF gotten by the delayed CSIT is bigger than N DoF when there is no CSIT. More exactly, KN/(1+1/2+...+1/K) DoF can be achieved in the MIMO BC where the transmitter is equipped with KN antennas and there are K receivers with each one equipped with N antennas;6N/5DoF is achievable in MIMO X channel where each node equipped with N antennas interference channel with3-users can reach9N/8DoF.