Wavelet Transform and GRU-Enhanced Deep Reinforcement Learning for Adaptive Bitrate Control in Video Streaming

With the complexity of the network environment, the traditional video streaming transmission method faces the challenges of bandwidth fluctuation and delay variation, which seriously affects the user's quality of experience (QoE). In this paper, an adaptive bitrate (ABR) control method based on augmented deep reinforcement learning is proposed, which obtains multi-scale timing features by wavelet transform, and at the same time effectively models the timing features by using GRU, and finally dynamically adjusts the bitrate strategy by using PPO algorithm to optimize the user experience. Experimental results show that compared with the existing algorithms, the proposed WTGRU-DRL method improves the QoE by an average of 5%-60% under bandwidth and delay variations, which significantly improves the stability of video streaming and viewing experience.