Interference management techniques with imperfect channel state information at the transmitter

Abstract

Interference is the source of the most serious performance impairment in today's wireless communication networks. Recent research results have highlighted the importance of interference coordination in such networks. There are several schemes that effectively manage the interference assuming that the state of the channel is known at the transmitters. However, having access to perfect channel state information (CSI) at the transmitters is not a realistic assumption. The aim of this dissertation is to study and develop methods enabling interference coordination in a wireless network while having imperfect channel state information at the transmitters. In the first part of this thesis, advanced channel state representations are employed in order to cope with the problem of interference when the transmit signals are designed based on imperfect CSI available at the transmitter. Efficient quantization of the CSI is investigated to reduce the requirement for information exchange over the network and in particular feedback to the transmitters. Different scenarios are considered where availability of CSI at the transmitter is crucial to achieve high throughput. In the second part of this thesis, a particular type of CSI imperfection is considered where the available CSI at the transmitter is completely outdated with respect to the current state of the channel. A simple method is proposed to exploit the outdated CSI in a multiple-input multiple-output (MIMO) two-user Gaussian interference channel (IC). The proposed scheme is shown to achieve the optimal degrees of freedom (DoF) of this channel. In the third part of this thesis, it is assumed that the transmitters have access to local CSI and the process of designing the transmit signal is distributed over the network. A message-passing framework is proposed to effectively model the information exchange over the network when the goal is to obtain an interference alignment (IA) solution in a distributed manner. In the last part of this thesis, the uncertainty about the channels at the transmitters is modeled as an independent additive Gaussian error. This simplifies the performance analysis and allows for the optimization of the transmit signal to ensure robustness against channel uncertainty and obtain solutions that are adaptive to the channel condition. Two different approaches are proposed to minimize the impact of the residual interference caused by the channel uncertainty.