Q-learning for POMDP: An application to learning locomotion gaits

Keyphrases

• Gait 100%
• Partially Observable Markov Decision Process 100%
• Q-learning 100%
• Locomotion Gait 100%
• Learning Locomotion 100%
• Posterior Distribution 50%
• Learning Problems 50%
• Learning Framework 50%
• Main Components 50%
• Control Architecture 50%
• Dynamic Model 50%
• Rigid Body 50%
• Simulation Environment 50%
• Coupled Oscillators 50%
• Noisy Measurements 50%
• Robotic System 50%
• Optimal Control 50%
• Phase Variable 50%
• Optimal Control Law 50%
• Q-learning Algorithm 50%
• Sensory Measurement 50%
• Control Sets 50%
• Periodic Input 50%
• Phase Modeling 50%
• Periodic Gait 50%
• Feedback Particle Filter 50%
• Numerical Optimal Control 50%
• Two-body System 50%
• Optimal Locomotion 50%
• Biolocomotion 50%

Engineering

• Q-Learning 100%
• Partially Observable Markov Decision Process 100%
• Optimal Control 66%
• Open Loop 33%
• Objective Function 33%
• Rigid Structure 33%
• Main Component 33%
• Control Architecture 33%
• Learning Algorithm 33%
• Posterior Distribution 33%
• Particle Filter 33%
• Robotic System 33%
• Simulation Environment 33%
• Phase Modeling 33%
• Control Law 33%
• Oscillator 33%