Nonlinear Digital Self-Interference Cancellation with Reduced Complexity for Full Duplex Systems

Full duplex transmission is currently viewed as an important technology component for the future 5G and beyond mobile broadband technology. In order to realize its promised theoretical gain, sufficient cancellation of the self-interference must be achieved. The focus throughout this work will be on the digital cancellation, which main target is to cancel the residual self-interference resulting from the insufficient analog cancellation due to hardware imperfections, along with non-linearities existing in the transmitter chain. A novel pre-transmission transformation based on the Cholesky decomposition is presented, that aims at enhancing the digital cancellation performance. A digital cancellation based on the transversal recursive least squares with the exploitation of the dichotomous coordinate descent algorithm to lower the computational complexity is presented. The analysis was extended to include the existence of a received signal of interest, while simultaneously canceling the self-interference signal. By means of numerical simulations, a performance evaluation was carried out and results showed that the level of residual interference after the digital canceler reaches the simulated noise floor power level.