PDCF-DRL: A Contention Window Backoff Scheme Based on Deep Reinforcement Learning for Differentiating Access Categories

Abstract Purpose: In wireless network backoff algorithms, the application of Deep Reinforcement Learning (DRL) allows the agent to learn through interactions with the environment, enabling dynamic adjustment of the contention window (CW) size to improve network performance and throughput. However, existing wireless network backoff algorithms integrated with DRL still face problems such as inappropriate backoff strategies, high latency during the backoff process, suboptimal throughput under heavy load conditions, and failure to satisfy the need for differentiating access categories (ACs).To further enhance the performance of wireless networks and meet the various quality of service (QoS) requirements of users based on AC differentiation. Methods: This paper proposes a CW backoff scheme based on DRL for differentiating ACs. The scheme uses DRL technology to observe the channel collision rate and sense the current network conditions, thereby adaptively adjusting the CW size for ACs. In addition, it dynamically adjusts the backoff strategy according to the perceived current network conditions to optimize the data transmission process. Results: When the number of competing stations increases from 20 to 120, with an equal number of stations for each AC, the overall throughput of this scheme ranges from 82.86% to 80.81%, representing an improvement of 34.2% to 72.1% over the traditional EDCA mechanism. The overall collision rate ranges from 2.36% to 9.15%, which is a reduction of 43% to 79.18% compared to the traditional EDCA mechanism. Additionally, there is a significant improvement compared to existing deep reinforcement learning-based optimization schemes. Conclusion: Experimental results show that this scheme effectively discriminates between different ACs, resulting in lower latency and higher throughput for high-priority traffic. Furthermore, adaptively adjusting the CW size and improving the backoff strategy maintains a low collision rate and stable throughput even under heavy load conditions, significantly improving the overall network performance.