Collision Avoidance in IEEE 802.11 DCF using a Reinforcement Learning Method

In IEEE 802.11 networks, wireless stations try to avoid congestions using the random backoff algorithm in distributed coordination function (DCF); however, for a large number of nodes, the network performance can be degraded due to increased collision probabilities. In this paper, we propose a new backoff algorithm that enables the stations to select backoff times such that no collisions occur as follows: Time is divided into frames, and each frame consists of a fixed number of time slots. Then, using a reinforcement learning method, each station selects its own time slot(s), which corresponds to its backoff time(s). Since every station has its own time slots to start with the same value of the backoff counter, collisions can be avoided and the network can achieve high performance. Our simulation results show that the proposed algorithm outperforms the IEEE 802.11 DCF method in various network environments.