Vision-based reinforcement learning for humanoid behavior generation with rhythmic walking parameters.
Masaki OginoYutaka KatohMasahiro AonoMinoru AsadaKoh HosodaPublished in: IROS (2003)
Keyphrases
- reinforcement learning
- humanoid robot
- function approximation
- maximum likelihood
- augmented reality
- parameter estimation
- human computer interaction
- expectation maximization
- real time
- real robot
- parameter settings
- parameter space
- human gait
- vision system
- markov random field
- learning process
- computer vision
- learning algorithm
- neural network