Improved Grant-Free Access for URLLC via Multi-Tier-Driven Computing: Network-Load Learning, Prediction, and Resource Allocation

Grant-Free (GF) access has been recognized as a promising candidate for Ultra-Reliable and Low-Latency Communications (URLLC). However, even with GF access, URLLC still may not effectively gain high reliability millimeter-level latency, simultaneously. This is because the network load typically time-varying known to base station (BS), thus, resource allocated cannot well adapt variations of load, resulting in low utilization efficiency under light leading severe collisions heavy load. To tackle this problem, we propose multi-tier-driven computing framework associated algorithms support users different QoS requirements. Especially, concentrate on $K $ - repetition its simplicity well-balanced performance practical systems. In particular, our consists three tiers computation, namely network-load learning, prediction, adaptive allocation. first tier, BS can learn information from states (success, collision, idle) random-access resources terms blocks (RB) time slots. second variation predicted based estimation results tier. Finally, third by deriving weighing failure probabilities groups users, their requirements, loads, able dynamically allocate sufficient accommodating varying loads. Simulation show that proposed approach estimate more accurately compared baseline schemes. Moreover, assistance allocation offers an effective way enhance services,