Secure Satellite Downlink with Hybrid RIS and AI-Based Optimization

• Published in: IEEE access 2024-12, p.1-1
• Main Authors: Kamal, Mian Muhammad, Abideen, Syed Zain Ul, Shah, Syed Sikandar, Sehito, Nabila, Khan, Salahuddin, Virdee, Bal S., Alibakhshikenari, Mohammad, Livreri, Patrizia
• Format: Article
• Language: English
• Subjects: Active and passive RIS elements; Deep reinforcement learning (DRL); Hybrid reconfigurable intelligent surface (RIS); Physical Layer Security (PLS)
• ISSN: 2169-3536
• DOI: 10.1109/ACCESS.2024.3520796

In this paper, we explore a secure multiuser multiple-input single-output (MISO) satellite downlink communication system, enhanced by the integration of a hybrid reconfigurable intelligent surface (RIS). The study formulates a robust joint design for satellite and RIS beamforming, aimed at maximizing the secrecy rate of the overall system. Both the active and passive elements of the RIS are optimized, taking into account practical models that reflect real-world constraints, such as outdated channel state information (CSI) and the power consumption of the system. To address the highly complex, dynamic, and multidimensional nature of the beamforming design problem, deep reinforcement learning (DRL) techniques are employed. Simulation results demonstrate the effectiveness of the proposed beamforming strategy, highlighting significant performance improvements when utilizing hybrid-RIS compared to traditional passive RIS solutions in wireless communication systems.