Modified Analog Least Mean Square Loop With Band-Pass Filter Bank for Self-Interference Cancellation in In-Band Full-Duplex Systems

This paper presents a modified analog least mean square (MALMS) loop designed for self-interference (SI) cancellation (SIC) in in-band full-duplex (IBFD) communication systems. The MALMS loop utilizes radio frequency (RF) tapping, eliminating the need for auxiliary chains or mixers for upconversion. Unlike traditional approaches that rely on tapped delay lines (TDLs), the proposed MALMS loop incorporates an analog bandlimited band-pass filter (BPF) bank, effectively addressing the limitations and dependencies associated with TDLbased structures. Comprehensive analyses are conducted on some classical filter banks' phase and group delay characteristics, focusing on Butterworth and Chebyshev topologies. Furthermore, the performance of SI channel modeling in the MALMS loop is evaluated using various filter banks across different tap numbers. Simulation results demonstrate that the MALMS loop employing the BPF bank can achieve superior SIC compared to conventional TDL adaptive filters if low complexity implementation is desired, providing an efficient and practical solution for IBFD communication systems.