Neurodynamics Adaptive Reward and Action for Hand-to-Eye Calibration With Deep Reinforcement Learning.
Zheng ZhengMengfei YuPengfei GuoDelu ZengPublished in: IEEE Access (2023)
Keyphrases
- reinforcement learning
- action selection
- gaze estimation
- reward shaping
- state action
- state space
- action space
- gaze detection
- temporal difference
- function approximation
- partially observable domains
- gaze tracking
- dynamic programming
- eye tracking
- reward function
- learning capabilities
- reward signal
- eye movements
- camera calibration
- reinforcement learning algorithms
- optimal policy
- agent learns
- model free
- gaze direction
- eligibility traces
- learning algorithm
- transition model
- calibration procedure
- hand movements
- control policy
- eye gaze tracking
- head movements
- eye gaze
- partially observable environments
- markov decision processes
- multi agent
- real time
- agent receives