Radio self-interference cancellation by transmit beamforming, all-analog cancellation and blind digital tuning

Radio self-interference cancellation has been a technological challenge for more than a century while it is the most critical enabler for full-duplex radios. The eventual success of radio self-interference cancellation may well depend on not only improved hardware technology but also innovative signal processing schemes. In this paper, we present a few latest discoveries on such schemes. The first is an improvement of time-domain transmit beamforming with robustness against the IQ imbalances in radio circuits, which is supported by both simulation and hardware experimental results. A key innovation here is due to the use of real-valued linear model instead of complex-valued linear (or widely linear) model. The second is a numerical investigation of the performance limits of an all-analog cancellation channel based on clustered-taps of attenuators when the interference channel has a large number of random multipaths. The third is a blind digital tuning method which uses only the baseband waveforms to determine the values of the variable attenuators embedded in the all-analog cancellation channel. This method is robust against imperfections in the knowledge of the transfer function of any component in the system provided that a real-valued linearity property holds (except for the transmit chain).