Practical Implementation Issues of Reinforcement Learning Based ALOHA for Wireless Sensor Networks

This paper presents a practical implementation of the Reinforcement Learning (RL) based ALOHA protocols (RLALOHA) for single-hop communication. The practical implementation issues of RL-ALOHA are studied. A potential real-world phenomenon that impacts on performance is brought to light and a new scheme is proposed to deal with the phenomenon which provides the ideal throughput performance of RL-ALOHA. The practical results show that the proposed scheme improves the convergence properties to the steady state of one slot assigned per node.