Adaptive MAC Scheduling Strategy Based on Channel Sensing and Reinforcement Decision in Distributed Internet of Underwater Things

Distributed Internet of Underwater Things (D-IoUT) has broad application prospects in marine monitoring, resource development, and security protection. However, the spatio-temporal uncertainty of underwater acoustic networks and the presence of multiple collision domains pose severe challenges to MAC-layer scheduling and interference management. To address these issues, this paper proposes an adaptive scheduling MAC protocol based on channel sensing and reinforcement decision-making (CSRD-ASMAC), aiming to provide an efficient and adaptive transmission scheduling scheme for D-IoUT. The protocol first introduces a hierarchical channel-state classification framework to accurately characterize the channel environment for each slot. Building on pre-divided basic slots, it integrates channel sensing and slot prediction into a reinforcement learning framework and employs a mixed-action exploration strategy to update Q-values, enabling each node to adaptively select the optimal transmission slot. Simulation results show that CSRD-ASMAC effectively improves overall network throughput in multi-collision domain environments.