Channel Access Control Instead of Random Backoff Algorithm

In wireless communication systems that perform carrier sense, packet collisions due to simultaneous transmission between transmitting nodes frequently occur because transmission starts the moment the channel becomes idle. In wireless LANs and other systems, backoff algorithms are used to avoid simultaneous transmission, but the commonly used binary backoff results in excessively large waiting times due to random backoff. Therefore, this paper proposes a new channel access control method using reinforcement learning. Simulation evaluation shows the effectiveness of the proposed method by the characteristics of the transmission success rate.