The Key Technologies Of Interference Alignment Precoder Design

Interference Alignment as a novel interference processing method, has attracted the extensive attention of academia in recent years. Its key idea is to design the transmit signals with specific structure so that the unwanted signals from multiple interferers are aligned into particular subspace with reduced dimensions at the receivers. It can achieve much higher wireless networks capacity than previously believed. Therefore interference alignment will contribute to the reliable transmission of large capacity in the case of multi-user interference; on the other hand, it can promote the development of network information theory, especially play a pivotal role in the study of multi-user multi-antenna interference network regional capacity boundaries. The main way to achieve interference alignment is employing the precoder technology. Therefore the application of interference alignment precoder design technology is significant in practical communication systemThe main idea of this dissertation is to fully understand the physical meaning of interference alignment and the degrees of freedom, propose the key issues of interference alignment precoder design, and provide useful solutions.Interference alignment precoder design based on the characteristics of high-dimensional matrix:we focus on the multi-user multi-input multi-output (MU-MIMO) system by limiting the optimization only at the transmitters ’side. To overcome the limitations of previous research, we relax the assumption of channel reciprocity and propose the algorithm based on the distance between matrix and space. It bypasses the overhead generated by alternation between the forward and reverse network, and it is applicable to both TDD and FDD systems. Moreover, different from the traditional transceiver jointly design method, we propose the idea of transmitter unilateral preoder design by employing the algorithm based on distance between interference space. Then by minimizing the remnant interference in the useful signal space, we provide the steepest decent algorithm to design the interference alignment precoder. More importantly, the simulation results suggest that the steepest decent method converges faster and has smaller remnant interference than other methods. The angles between interference space further evidence that the algorithm can indeed achieve interference alignment. Interference alignment precoder design based on manifold optimization theory: mainly to solve the key issues of optimization theory and methods for interference alignment precoder design. We introduce the novel manifold optimization theory into the interference alignment precoder design for MU-MIMO system. After a short introduction about manifold optimization theory, we reform our problem on Stiefel manifold, then derive the first order partial derivatives and the gradient direction of Stiefel manifold. Finally by employing the manifold geodesic, we propose the steepest descent method on Stiefel manifold for interference alignment precoder design. Moreover, we explore the unitary invariance property of our cost function, and discuss the relationship between Stiefel and Grassmann manifolds, to solve the optimization problem on Grassmann manifold. Through simulation and theoretical analysis, we find that the algorithms based on manifold optimization has lower complexity and faster convergence than traditional methods.The precoder design based on subspace interference alignment:mainly to solve the key issues of subspace interference alignment precoder design for Interference Multiple Access Channel and Interference Broadcast Channel of cellular system. By employing subspace interference alignment, we decompose the total space as the product of some subspaces at each receiver. To align the interference from different users into corresponding subspaces, we build the object function by using subspace interference covariance. Meanwhile because of the increment of interference sources, we modify the steepest descent method on Grassmann manifold by introducing the tangential space and projection operation. It will alleviate the redundant complexity generated by manifold geodesic; more significantly, we can use adaptive step choosing rule in the tangential space of Grassmann manifold. Through simulation, we find the proposed method has higher capacity and faster convergence than previous methods. Finally we deeply analyse channel decomposability and link duality.In all the proposed issues are significant to the application of interference alignment in the practical communication system, and the proposed solutions can effectively improve the performance of multi-user interference network.