SCMA-Q-learning with overload control for random access in LEO satellite mMTC networks

Abstract

With the continuous advancement of 5G and beyond wireless networks, the exponential increase in Machine-Type Communication (MTC) devices presents substantial challenges, particularly in deal with Random Access Channel (RACH) congestion in massive MTC (mMTC) networks. Low Earth Orbit (LEO) satellite communication networks are increasingly valued for their global coverage and low-latency benefits. This paper deals with the RACH congestion problem in LEO satellite networks by proposing an innovative method that combines Sparse Code Multiple Access (SCMA), Q-learning, and Access Class Barring (ACB) technology. The proposed approach allows MTC devices to dynamically optimize resource allocation by selecting the most appropriate codebooks and time-slot groups in response to environmental changes. Additionally, ACB is employed to control system overload by adjusting the access probability of devices based on network conditions. Simulation results show that the proposed SCMA-QL-ACB method significantly improves system throughput and reduces collisions compared to existing techniques.