Trajectory Planning With Deep Reinforcement Learning in High-Level Action Spaces.
Kyle R. WilliamsRachel SchlossmanDaniel WhittenJoe IngramSrideep MusuvathyJames PaganKyle A. WilliamsSam GreenAnirudh PatelAnirban MazumdarJulie ParishPublished in: IEEE Trans. Aerosp. Electron. Syst. (2023)
Keyphrases
- action space
- trajectory planning
- reinforcement learning
- state space
- state and action spaces
- motion planning
- continuous state
- markov decision processes
- robot manipulators
- real valued
- obstacle avoidance
- dynamic environments
- path planning
- action selection
- skill learning
- single agent
- reinforcement learning algorithms
- function approximators
- stochastic processes
- degrees of freedom
- multi agent
- markov decision process
- state action
- mobile robot
- learning algorithm
- continuous state spaces
- situational awareness
- markov decision problems
- function approximation
- partially observable
- optimal policy
- policy search
- model free
- continuous action
- reward function