Study On Algorithms Of Interference Alignment In Wireless Communication Systems

With the rapid development of wireless communication technology, interferencealignment (IA) technique has been the research focus in multi-user wirelesscommunication. IA enables multiple transmitters simultaneously communicate with theirparied receivers without interfering each other. This technique can align the signal from theother users which is made as interference signal for the desired user to an unused subspaceof the same signal space called interference subspace. Since interference subspaceseperates from the desired signal subspace, interference for the desired user is decreased.IA technology maximizes the DoF of the system. Most IA schemes are proposed in MIMOchannels that the combination of IA and MIMO technique can improve system’s capacitysignificantly. This paper solves the problems of existing interference alignment algorithmsbased on multi-user MIMO wireless communication system in order to improve the systemperformance.Multi-user wireless communication system includes cognitive multi-user wirelesscommunication system and traditional multi-user wireless communication system. We willpropose new IA algorithms in this two wireless communication system respectively.1. Cooperative Interference Alignment Max-SINR algorithmIn the cognitive radio system, Cooperative Interference Alignment (CIA) is that theprimary receiver performs interference suppression filtering participating in IA, buttransmitter does not. The primary filter eliminates interference from second users throughcompressing interference. The classical algorithm of CIA is Cooperative DistributedInterference Alignment (CDIA) algorithm. CDIA algorithm is an iterative procedure thatupdates system’s filters utilizing the reciprocity of alignment to decrease the totalinterference leakage progressively. If the leakage interference converges to zero, theninterference alignment is feasible. However, this algorithm just realizes IA and no coherentcombining gain for the desired signal is obtained with IA.In order to solve the above problems, this paper provides a new algorithm calledcooperative interference alignment Max-SINR (CIAMSINR) algorithm which maximizesSINR on the basis of the primary user cooperative interference alignment (CIA).CIAMSINR algorithm utilizes the duality of channel and updates pre-coding matrix andinterference compressed matrix in the forward channel and dual channel so that the signalto interference noise ratio reaches a maximum. Thus, the proposed algorithm completes IA.In the system whose DoF allocation is already determined, the propose IA realizesinterference alignment as well as improves users’ transmission rates. Simulation results show the effectiveness of our CIAMSINR and its performance in the transmission rates.2. Robust MMSE interference alignment algorithm on non-ideal MIMO X channelMost IA cases are realized on the ideal channel. However in the actual communicationsystem, it’s difficult to get perfect CSI and the channel is non-ideal. The performance of IAalgorithms based on perfect CSI drop dramatically in circumstance where CSI errors exist.In consideration of this problem, According to the nature of interference alignment (IA), inconsideration of imperfect channel state information (CSI) of MIMO X channel, a robustminimum mean square error interference alignment (robust-MMSE-IA) algorithm usingMSE objective function and based on the non-ideal MIMO X channel l is proposed. Thisalgorithm uses symbol extension and MMSE criterion so that the objective function isconvex after deciding pre-coding matrix to optimize interference compression matrix orvice versa. As a consequence, the proposed algorithm can be converted to two convexsub-problems, and optimal pre-coding matrix and interference compression matrix can beobtained respectively. Simulation for the proposed robust-MMSE-IA algorithm andtraditional non-robust-MMSE-IA algorithm is provided on non-ideal channel to showperformance improvement of the proposed algorithm in interference alignment undernon-ideal MIMO X channel.