Enhancing Wi-Fi in Dynamic Interference Environments: A Joint Rate Adaptation and Channel Access Approach via Deep Reinforcement Learning

The growing density of Wi-Fi access points (APs) and devices has created interference-prone network environments, leading to degraded and unstable Wi-Fi performance. This paper introduces JRACA, a novel joint rate adaptation and channel access scheme based on deep reinforcement learning, designed to maximize Wi-Fi throughput in dynamic interference scenarios. JRACA leverages the Proximal Policy Optimization (PPO) algorithm and models the problem as a Markov Decision Process (MDP). By observing states such as the Signal-to-Noise Ratio (SNR) of the acknowledgement (ACK) frame, the last Modulation and Coding Scheme (MCS), the backoff window statistics, and transmission success rates, JRACA dynamically adjusts MCS, clear channel assessment (CCA) threshold, and backoff window sizes. Extensive simulations experiments are conducted using the open-source ns- 3 simulator, showing that JRACA outperforms both MinstrelHt and an RL-based baseline scheme, achieving up to $40.02 \%$ and $44.81 \%$ higher throughput, respectively. Furthermore, JRACA exhibits robust generalization capabilities in scenarios that differ from its training environment.