Deep Reinforcement Learning With Optimized Reward Functions for Robotic Trajectory Planning.
Jiexin XieZhenzhou ShaoYue LiYong GuanJindong TanPublished in: IEEE Access (2019)
Keyphrases
- reward function
- trajectory planning
- reinforcement learning
- mobile robot
- obstacle avoidance
- motion planning
- reinforcement learning algorithms
- markov decision processes
- policy search
- state space
- path planning
- robot manipulators
- optimal policy
- inverse reinforcement learning
- dynamic environments
- markov decision process
- transition model
- function approximation
- transition probabilities
- real time
- robotic systems
- state variables
- multiple agents
- dynamic programming
- multi agent
- neural network
- machine learning
- learning algorithm
- decision making
- continuous state
- state action
- generative model
- degrees of freedom
- model free
- autonomous robots
- action space
- markov chain