On the Impacts of I/Q Imbalance in Analog Least Mean Square Adaptive Filter for Self-Interference Cancellation in Full-Duplex Radios

Frequency-dependent in-phase/quadrature-phase imbalance (IQI) is a critical hardware impairment, which affects the performance of the self-interference (SI) cancellation structure in full-duplex systems. A novel frequency-dependent IQI model is proposed in this paper, providing a theoretical tool to evaluate the impacts of this impairment in the analog domain. The model is then applied to investigate the performance of the analog least mean square (ALMS) loop under both frequency-selective and frequency-independent IQI. The degradation in the SI cancellation performance of the ALMS loop is revealed and its upper bound is derived. Numerical results show that the ALMS loop is resilient to both types of IQI with the maximum level of only 3.5 dB performance degradation even when the image rejection ratio is just 15 dB.