URLLC for DRL-RIS-AIDED SAGINs: New Mentalities, Trends and Preliminary Solutions

Future communication networks will seamlessly integrate satellite, aerial platforms, terrestrial and mar-itime communications systems to support ultra-re-liable and low-latency communication (URLLC) services. The next generation of wireless communication is trending towards green, efficient and real-time, hence the recent emergency of reconfigurable intelligent surface (RIS) has been recognized as a pathway to accelerate the realisation of ubiqui-tous connectivity. This is because RIS can bring the advantages of high energy efficiency, low latency and no additional decoding compared to traditional relay solutions. Motivated by this innovative technology, firstly, this article comprehensively discusses new mentalities and trends in RIS-assisted space-ae-rial-ground integrated networks (SAGINs). Then, we identify potential challenges in meeting connectivity requirements and examine key techniques for optimising network performance by using deep rein-forcement learning (DRL). In addition, we present some representative cases of the possibility of inte-grating RIS into SAGINs URLLC system using various multiple access (MA) techniques. Furthermore, sev-eral simulation results are provided to demonstrate the advantages of the proposed DRL-enabled proximal policy optimization (PPO) algorithm in terms of system performance improvement.