Model-Driven Deep Learning for Joint Waveform and Beamforming Design in RIS-ISAC Systems

Integrated sensing and communication (ISAC) has become a crucial technology in future wireless systems. The recently emerged symbol-level precoding (SLP) technique is promising for ISAC waveform design since it can provide better multi-user communication and radar sensing performance by leveraging both temporal and spatial design degrees of freedom (DoFs). Previous research has been confronted with the huge computation burden challenge brought by the SLP design. In this paper, we propose a novel model-driven deep learning based SLP design by unfolding an iterative alternative direction method of multiplier (ADMM) algorithm, in order to efficiently solve the joint ISAC waveform and reconfigurable intelligent surface (RIS) beamforming design problem for RIS assisted ISAC systems. Specifically, our goal is to maximize the radar output signal-to-interference-plus-noise ratio (SINR) while satisfying the communication quality-of-service (QoS) requirements. Simulation results verify that our proposed model-driven deep learning algorithm achieves comparable radar sensing performance to the prior optimization-based approaches, while providing a noteworthy reduction in computational complexity in terms of average execution time.