Artificial-Intelligence-Empowered External Interference Cancellation Using Reconfigurable Intelligent Surfaces for In-Band-Full-Duplex Systems

In this article, an artificial-intelligence-empowered external interference cancellation method utilizing reconfigurable intelligent surfaces (RISs) for in-band full-duplex (IBFD) systems is proposed. Two kinds of representative external scatterers are introduced into a miniaturized dual-polarized full-duplex system with two collocated transmitter (Tx) and receiver (Rx) ports, thus causing a severe deterioration in the isolation between Tx and Rx ports. With the help of the miniaturized dual-polarized bandpass RIS, the transmission phase of the inherent space coupling waves is regulated to neutralize the couplings from external interferences. The transmission phases of RIS are optimized offline based on a greedy strategy, where artificial neural networks (ANNs) based on measurement datasets are trained as the surrogate model for describing the function of transmission phases of RIS and coupling of Tx and Rx ports, and invoked in the optimization process. Five illustrative experiments are performed to demonstrate the effectiveness of the system with respect to different external interferences with distinct forms and time delays, as well as different phase resolutions of RIS. The average and worst isolation between Tx and Rx ports are significantly enhanced by around 15 dB within the 200 MHz band in these conditions. The proposed method has the potential to be applied to future 6G wireless systems with full-duplex operation mode.