An Access Control Scheme Combining Q-Learning and Compressive Random Access for Satellite IoT

The satellite Internet of Things (IoT) covers a vast area with multiple nodes and has limited resources for random access, which leads to low throughput. In this letter, we propose a Q-learning combined with diversity slotted compressive random access control scheme (QDCC) to enhance resource utilization and throughput. Moreover, we address the issue of reduced throughput during network overload by introducing an adaptive frame length QDCC (AFL-QDCC) scheme. This scheme adjusts the frame length by utilizing the support sets estimated by the compressive sensing (CS) reconstruction algorithm. Simulation results demonstrate that the proposed QDCC scheme outperforms conventional schemes in terms of throughput. Furthermore, the AFL-QDCC scheme can maintain stable and high throughput performance even with a large number of nodes.