Reinforcement Learning endowed with safe veto policies to learn the control of Linked-Multicomponent Robotic Systems.
Borja Fernández-GaunaManuel GrañaJosé Manuel López-GuedeIsmael Etxeberria AgirianoIgor AnsoateguiPublished in: Inf. Sci. (2015)
Keyphrases
- robotic systems
- reinforcement learning
- imitation learning
- optimal policy
- autonomous robots
- control architecture
- reinforcement learning agents
- mobile robot
- real robot
- vision system
- control policies
- hierarchical reinforcement learning
- control policy
- robotic manipulator
- control problems
- robot control
- search and rescue
- manipulation tasks
- object manipulation
- multi robot
- robot behavior
- state space
- optimal control
- unstructured environments
- policy search
- markov decision process
- machine learning
- function approximators
- markov decision processes
- transfer learning
- multi agent
- real time
- motor control
- action selection
- control strategies
- robot navigation
- particle swarm optimization
- computer vision
- learning algorithm
- human teaching behavior
- impedance control