The ASL360 system introduces a novel approach to enhance the streaming experience of 360 degree video by employing adaptive deep reinforcement learning techniques. By utilizing a Constrained Markov Decision Process (CMDP), ASL360 intelligently schedules video segments to optimize quality while managing network conditions. The system dynamically adjusts its policies based on real-time data, ensuring that users receive the best possible experience despite fluctuations in network performance.
A significant focus of ASL360 is on improving critical QoE metrics, such as video quality and rebuffering times. In experimental results, ASL360 demonstrated remarkable enhancements, achieving approximately 2 dB higher video quality with an 80% reduction in average rebuffering time compared to existing methods. This progress underscores its potential to advance immersive video experiences, making it particularly valuable in dynamic environments where network reliability can vary dramatically.
👉 Pročitaj original: arXiv AI Papers
