Optimal Asynchronous Cooperative Random Access in Satellite-Enabled Industrial Internet of Things

The satellite-enabled Industrial Internet of Things (S-IIoT) plays a crucial role in supporting uplink access of terminals in remote areas where terrestrial networks are unavailable. However, the massive access of IIoT terminals and their low-power limitations have greatly reduced the system performance of random access (RA) in S-IIoT. To address this issue, we propose an optimal asynchronous cooperative random access (OACRA) for uplink access in S-IIoT, which greatly enhances the system performance with the aid of cooperative relays. Specifically, a cooperative transmission optimization problem is formulated to find out multiple optimal cooperative relays for each user. To address this problem more efficiently, a reinforcement learning based distributed solution is proposed to obtain the optimal solution in a distributed way, which is able to capture dynamic features of moving satellite in real-time. We then propose the OACRA protocol by jointly designing the packet transmission mode at the transmitter and multi-user detection at the receiver. At the transmitter, each user transmits multiple packet replicas to the satellite through direct link and relay links in an asynchronous way, which enables asynchronous packet diversity transmission with power-domain non-orthogonal multiple access (NOMA). At the receiver, a sliding window-based multi-user detection is proposed to detect the partially collided packet replicas in an iterative way. We further analyze theoretically the performance of the OACRA protocol, which provides some insights into theoretical analysis in unequal-power asynchronous case. Simulation results show that the proposed OACRA protocol significantly outperforms the state-of-the-art protocols in terms of normalized throughput, packet loss rate (PLR), normalized sum rate, and energy efficiency.