AI-Driven Integration of Sensing and Communication in the 6G Era

Attributing to the rapid growth of AI, the integration of sensing and communication (ISAC) networks has embraced AI in the upcoming new-style mobile communication networks. A FedFog network architecture for ISAC networks is proposed in this article, which consists of the terminal perception layer, the edge base station processing layer, and the cloud data layer. In the context of multiple base stations (BSs), the handover between BSs and user equipment is worthy to be studied. Referring to the concept of coordinated multiples BSs, we design a handover procedures in the ISAC networks. Meanwhile, a federated reinforcement learning scheme of user control is designed. However, due to new unlicensed spectrum bands such as millimeter wave band and Terahertz band, the hybrid beamforming can reduce the expenses of hardware. A learning-based interference management utilizing the hybrid beamforming is designed. Meanwhile, we consider self-interference and mutual interference cancellation with deep neural networks. Simulation results show the performance of AI-driven ISAC networks in terms of mobility and interference management, and further prove that services are boosted for 6G networks.