Adaptive Self-Interference Cancellation for Full Duplex Radio: Analytical Model and Experimental Validation

Recent experimental results show that full-duplex (FD) communication is feasible provided that the residual self-interference after cancellation is close to the noise floor. In this paper, we present a low-complexity active self-interference cancellation technique for FD orthogonal frequency division multiplexing (OFDM) systems, which accounts for the inherent cyclostationarity of the OFDM signal in order to effectively cancel the self-interference. The method here proposed estimates the self-interference cancellation path by using a time-averaged mean-square error criterion, combined with a filtered least mean squares adaptive algorithm. Then, its performance is experimentally validated using a self-developed FD radio testbed implemented with off-the-shelf components. When combined with passive cancellation, applying the proposed technique provides enough self-interference suppression as to reach close (less than 2 dB) to the receiver noise floor.