Interference Alignment And Its Application In Modern Wireless Communication Systems

Recently, with the rapid development of wireless communications, interferencemanagement becomes a key issue in modern wireless communication systems.Although many approaches have been proposed to address this issue, many basicproblems in wireless interference networks remain unresolved. Interference alignmenthas been introduced in the study of interference channels. It was a milestone in thestudy of interference management. Therefore, interference alignment and itsapplication in modern wireless communications will be studied in this dissertation, andthe main achievements and results of this dissertation are listed as follows:1. In the K-user single-input single-output frequency-selective fading interferencechannel, it is shown that the maximal achievable multiplexing gain is almost surely K/2by using interference alignment. However, when the signaling dimensions are limited,allocating all the resources to all users simultaneously is not optimal. So, a group basedinterference alignment (GIA) scheme is proposed, and it is formulated as anunbounded knapsack problem. Optimal and greedy search algorithms are proposed toobtain group patterns. Analysis and numerical results show that the GIA scheme canobtain a higher multiplexing gain when the resources are limited.2. The degrees of freedom (DoF) upper-bound is characterized for an L-cellconstant cellular interfering network by using linear interference alignment. For asymmetric network, it means that no matter how many cells and users per cell there are,the total achievable DoF is bounded by M+N, where M and N denote the number ofantennas equipped on users and base stations. Besides, a DoF achievable algorithm isalso proposed. Although the DoF upper-bound cannot be reached, more DoF can beobtained compared with orthogonal schemes. For a three-cell constant cellularinterfering network, a new property of alignment is identified, i.e., an interferencealignment solution obtained in a user-cooperation scenario can also be applied in anon-cooperation environment. By using this property, an ES-IA (Eigenvector Splitbased Interference Alignment) algorithm is proposed by jointly designing transmit andreceive beamforming matrices. Analysis and numerical results show that more DoF canbe achieved compared with conventional schemes in most cases. 3. Based on the interference alignment method, through cascading an inter-cellinterference rejection matrix before an intra-cell interference rejection matrix, the ratesof cell-edge users can be improved obviously. However, this method can only bedesigned for users at a fixed position while users are randomly distributed in realcellular systems. So, a dynamic mode switch (DMS) scheme is proposed. The users ina cell calculate a threshold according to its suffered interference and choose thepreferred mode accordingly (IA-MMSE or MF scheme), and an opportunistic dispatchis performed at the base stations for two groups. Simulation results show that the DMSscheme can achieve the improved performance of the cell-edge users while it doesn’taffect the rates of the cell-center users. At the same time, the rates of the cell-edgeusers are further improved through the optimization of matrix. Besides, an interferencealignment based space-division hybrid (IA-SDH) scheme is proposed by joint designof transmit precoding matrices and receive beamforming matrices for a three-cellconstant cellular network, where users at cell-center work with a conventional schemewhile users at cell-edge utilize an IA scheme. Cell-edge users suffer no interferenceand cell-center users only affected by negligible inter-cell interference. Analysis andnumerical results show that more users can be served simultaneously than that of aconventional scheme.4. Based on closed-form interference alignment solutions, a low overheaddistributed interference alignment (LOIA) scheme is proposed for the K-usersingle-input single-output interference channel, and extension to multiple antennascenarios is also considered. Compared with the distributed iterative interferencealignment (DIIA) algorithm proposed by Gomadam et al., the overhead of our LOIAscheme is greatly reduced. Simulation results show that the DIIA algorithm is strictlysuboptimal compared with our LOIA algorithm in the overhead limited scenario.Besides, the reciprocity of wireless channel has not been made full use in the DIIAalgorithm. Through cascading a precoding matrix before an interference alignmentmatrix, a PA-DIIA (Power Allocation DIIA) algorithm is proposed based on the DIIAalgorithm where different streams are allocated with different powers. APA-Max-SINR (Power Allocation Max-SINR) algorithm is also designed based on theMax-SINR algorithm where different SINR streams are allocated different powers.Simulation results show that our designed algorithms achieve more rates especially inlow power regime. Through a comparison of two kinds of algorithms, it is also showsthat PA-DIIA algorithm should be chosen under multiple streams and high power sceneries.