Harnessing Reinforcement Learning for Neural Motion Planning.
Tom JurgensonAviv TamarPublished in: Robotics: Science and Systems (2019)
Keyphrases
- motion planning
- reinforcement learning
- fitted q iteration
- degrees of freedom
- mobile robot
- trajectory planning
- path planning
- humanoid robot
- robot arm
- network architecture
- autonomous mobile robot
- robotic arm
- robotic tasks
- neural network
- sensory inputs
- belief space
- mechanical systems
- multi robot
- control law
- collision free
- inverse kinematics
- biologically inspired
- state space
- obstacle avoidance
- learning algorithm
- model free
- action selection
- kinematic model
- multi agent
- markov decision processes
- machine learning
- manipulation tasks
- optimal policy
- configuration space
- real robot
- dynamic environments
- feature space
- optimal control
- robot control