Adapting Biped Locomotion to Sloped Environments - Combining Reinforcement Learning with Dynamical Systems.
João AndréCarlos TeixeiraCristina P. SantosLino A. CostaPublished in: J. Intell. Robotic Syst. (2015)
Keyphrases
- dynamical systems
- reinforcement learning
- state space
- reinforcement learning methods
- partially observable
- control law
- differential equations
- robotic systems
- nonlinear dynamical systems
- control theory
- qualitative simulation
- robot control
- dynamic systems
- function approximation
- phase space
- optimal control
- mobile robot
- partially observable markov decision processes
- nonlinear dynamics
- policy search methods
- predictive state representations
- dynamical behavior
- linear systems
- markov decision processes
- humanoid robot
- multi agent
- action space
- control strategy
- optimal policy
- dynamic environments
- learning algorithm