Acmece/rl-collision-avoidance
A PyTorch implementation of deep RL-based decentralized multi-robot collision avoidance using PPO and ROS.
Not currently ranked — collecting fresh signals.
star history
This repository implements a deep reinforcement learning approach for decentralized multi-robot collision avoidance. It uses Proximal Policy Optimization (PPO) to train policies that enable robots to navigate crowded environments without collisions. The system supports curriculum learning with two training stages, progressing from simple to complex scenarios, and integrates with the ROS/Stage simulation environment. Training uses MPI for distributed computation across multiple processes.
Frequently asked
- What is Acmece/rl-collision-avoidance?
- A PyTorch implementation of deep RL-based decentralized multi-robot collision avoidance using PPO and ROS.
- Is rl-collision-avoidance open source?
- Yes — Acmece/rl-collision-avoidance is an open-source project tracked on heatdrop.
- What language is rl-collision-avoidance written in?
- Acmece/rl-collision-avoidance is primarily written in Python.
- How popular is rl-collision-avoidance?
- Acmece/rl-collision-avoidance has 458 stars on GitHub.
- Where can I find rl-collision-avoidance?
- Acmece/rl-collision-avoidance is on GitHub at https://github.com/Acmece/rl-collision-avoidance.