Toward Autonomous Resource Management Architecture for 6G Satellite-Terrestrial Integrated Networks

Abstract

Different from all existing mobile communication systems, the sixth-generation (6G) mobile communication system is expected to realize satellite-terrestrial integrated networks (STINs) and ubiquitous artificial intelligence (AI), which promotes that resource management (RM) realizes higher autonomy facing heterogeneous and high dynamic STIN and the 6G diverse service requirements. However, in what form will AI be applied in STIN's RM to give full play to its capabilities? How to construct an AI-integrated STIN's RM architecture to endow RM with higher autonomy and greater flexibility? In this article, AI is applied to STIN in an endogenous form, i.e., STIN has the ability to perceive and process information on service demands and resource states, and STIN can realize continuous optimization of service performance under unmanned conditions to play the role of AI more effectively. Further, this article combines the characteristics of AI and STIN to design an AI-centric three-level closed-loop (Resource-Access-Service-Access-Resource) intelligent RM (RASAR) architecture for STIN. Specifically, the RASAR architecture abstracts the STIN's RM into three independently deployable management functions and achieves satellite-terrestrial and access-bearer integrated resource management by designing the interfaces and the service performance feedback mechanism with the help of STIN's transmission capability and computing power. In addition, the closed-loop implementation of RASAR architecture promotes the combination of AI self-training and self-optimization of all levels to provide a solution for STIN's autonomous RM. Finally, a case study is presented, followed by a discussion of open research issues that are essential for the RASAR architecture.