UC Riverside

With fast increasing demand of the wireless network, the current spectrum used for commercial wireless communication becomes very crowed. So it is critical to find an more efficient way to make the limited spectrum provide larger capacity and throughput. Full-duplex communication technology has caused much attention in the past ten years since it can double the spectrum efficiency theoretically. While the main challenge obstructing it promoting into the market is the self-interference problem in a full duplex system. This dissertation focus on the self-interference cancellation (SIC) theories. The RF impairments occurred in the practical full-duplex system will be discussed. Among them, the phase noise and I/Q imbalance are regarded as the bottleneck of the self-interference cancellation and a detailed analyzing will be included in this dissertation. The general self-interference cancellation methods can be divided into passive self-interference cancellation and active self-interference cancellation where the active cancellation can be further divided into digital cancellation, analog cancellation and hybrid cancellation. This dissertation will review the theories of the passive and active self-interference cancellation. For the analog cancellation, two models (quadratic model and affine model) will be explored to handle the I/Q imbalance and phase noise. Both of these two models are based on the blind tuning algorithm which has two procedures: training and optimizing. This dissertation will introduce the development of the algorithm. It contains the computer simulation results as well as the hardware experimental results to prove the validation of the proposed ideas.