RL-adaptive foveated video streaming for VR robot teleoperation: simulator, double-DQN agents, a statistically careful study of reward design (with an honest negative result), and a live Three.js demo.
threejs reinforcement-learning robotics simulation pytorch virtual-reality dqn video-streaming mobile-robots teleoperation reward-design foveated-rendering adaptive-bitrate foveated-streaming
-
Updated
Oct 29, 2026 - Python