Q-Learning-Augmented Grant-Free NOMA for URLLC

Grant-Free (GF) Non-Orthogonal Multiple Access (GF-NOMA) has emerged as a promising technology for 5G networks requiring Ultra-Reliable Low Latency Communications (URLLC). However, the grant-free nature of these transmissions can introduce significant interference, thereby, negatively affecting URLLC system performance. To address this challenge, this paper introduces a novel, distributed GF-NOMA-based Q-learning framework that aims to minimize network latency based on a developed Mean Opinion Score (MOS) of packet age, while also maintaining high transmission success rates. Real-time feedback from the gNodeB (gNB) is employed to assist Machine-Type Devices (MTDs) in making adaptive decisions of joint power control and sub-carrier selection. Simulation results validate the effectiveness of our approach in minimizing delay and optimizing overall system performance.