A survey on AI-empowered RIS-assisted integrated sensing and communication: From terrestrial to non-terrestrial networks

Abstract

The integration of artificial intelligence (AI), reconfigurable intelligent surfaces (RIS), and integrated sensing and communication (ISAC) is emerging as a key enabler for intelligent, adaptive, and efficient wireless networks in the 6G era. This survey provides a comprehensive and well-structured overview of how AI is revolutionizing RIS-assisted ISAC by enabling dynamic, data-driven control over the radio environment. Through intelligent configuration of RIS, AI facilitates real-time adaptation of signal propagation paths, enhances sensing resolution, and ensures robust communication performance across diverse and dynamic network conditions. The discussion is organized across two major domains: terrestrial networks (TNs) and non-terrestrial networks (NTNs). In TNs, we examine four critical directions: (i) improving EE, (ii) strengthening security, (iii) achieving high sensing accuracy and low communication latency via joint optimization, and (iv) maximizing throughput. In NTN domain, especially within UAV-based platforms, we explore AI-RIS frameworks that support coordinated sensing and communication, reinforce link security in mobile and unpredictable environments, and enhance aerial network capacity through real-time RIS tuning. By bridging both terrestrial and aerial applications, this survey highlights how AI-driven RIS control enables environment-aware, low-latency, and secure ISAC operations. Finally, we identify and analyze key challenges such as managing trade-offs between sensing and communication, generalizing AI models across varied network scenarios, and ensuring scalable, low-complexity real-time deployment. The work concludes with forward-looking insights into future research directions essential for realizing robust and intelligent RIS-assisted ISAC architectures in 6G and beyond.