Skip to content

Generative AI-Driven Collaborative Beamforming and Trajectory Design in a Multi-UAV-RHS-Aided ISAC Secure System With Imperfect CSI

2026 · IEEE Transactions on Wireless Communications · Vol 25, pp. 26242-26255 · 0 citations · 35 references

Abstract

This paper investigates secure transmission in a multiple uncrewed aerial vehicle reconfigurable holographic surface (UAV-RHS)-aided integrated sensing and communication (ISAC) system, where multiple UAVs orchestrate simultaneous cooperative sensing and secure communication. We jointly design the UAVs’ hybrid RHS beamforming and trajectories along with UAV-user association, accounting for the imperfectly known channel state information (CSI) of the eavesdropper. A robust multi-objective optimization problem is formulated to maximize the sum secrecy rate and minimize the cooperative sensing Cramer-Rao lower bound (CRLB), subject to the constraints on the UAV swarm’s trajectories, the RHS radiation amplitude, and the total power budget. To address the NP-hard and mixed-integer dynamic non-convex problem, we propose a novel diffusion policy-enabled distributional soft actor-critic (DP-DSAC) approach, which is specifically designed for hybrid discrete-continuous action spaces. The innovative leverage of DP model for determining user scheduling and multi-UAV positioning allows the method to capture the complex dynamics and performance trade-offs in secure ISAC, leading to promising solutions. Simulation results demonstrate that the proposed DP-DSAC algorithm achieves 51% and 43% improvements in secrecy rate and positioning accuracy compared with non-diffusion learning-based methods.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.