Multi-agent Soft Actor-Critic Based Hybrid Motion Planner for Mobile Robots.
Zichen HeLu DongChunwei SongChangyin SunPublished in: CoRR (2021)
Keyphrases
- mobile robot
- multi agent
- actor critic
- motion planning
- reinforcement learning
- image sequences
- path planning
- multi agent systems
- dynamic environments
- policy gradient
- multi robot
- optimal control
- approximate dynamic programming
- optical flow
- autonomous robots
- heuristic search
- temporal difference
- policy iteration
- function approximation
- markov decision processes
- domain independent
- humanoid robot
- gradient method
- reinforcement learning algorithms
- model free
- neuro fuzzy
- state space