Two-Timescale Transmit Power Minimization in Semi-Passive IRS-Aided ISAC Systems

Passive intelligent reflecting surfaces (IRS) have become prevalent in integrated sensing and communication (ISAC) systems, creating a new link for sensing and communication for better performance without additional energy consumption. However, the echo signal will experience significant pathloss due to the passive IRS's high pathloss effect after multiple reflections, which limits the gain brought by IRS. To address this issue, this paper applies semi-passive IRS to mitigate severe pathloss and investigates its impact on the performance of ISAC systems. Leveraging the ability of semi-passive IRS to acquire statistical channel state information (CSI), we introduce a two-timescale design algorithm which employs stochastic successive convex approximation (SSCA) and semi-definite relaxation (SDR) to manage non-convex constraints. Simulation results demonstrate that the proposed two-timescale design achieves comparable transmit power to that of designs based on instantaneous CSI, validating the effectiveness and feasibility of the two-timescale design approach in ISAC systems.