Generalized Analog Least Mean Square Loop for Self-Interference Cancellation in In-Band Full-Duplex Communications

This paper introduces a novel method for radio frequency (RF) self-interference (SI) cancellation in in-band full-duplex (IBFD) communication systems using a generalized analog least mean square (GALMS) loop. Conventional research in this area has predominantly focused on employing the tapped delay line (DL) structure to process the RF transmitted signal. The proposed GALMS loop utilizes a filter bank applied on the baseband transmitted signal to avoid the limitation of the conventional tapped DL adaptive filters, which require tapped spacing to satisfy the Nyquist principle. We utilize frequency domain representation and steady-state analysis to derive the modeling error and residual SI power of the proposed structure. Notably, the GALMS loop incorporating an allpass filter bank can achieve the same level of SI cancellation as the conventional DL structure with a significantly reduced number of taps. Simulations are conducted to validate the efficiency of the proposed GALMS loop and prove the theoretical findings.