Extending the OpenAI Gym for robotics: a toolkit for reinforcement learning using ROS and Gazebo

Abstract

This paper presents an extension of the OpenAI Gym for robotics using the Robot Operating System (ROS) and the Gazebo simulator. The content discusses the software architecture proposed and the results obtained by using two Reinforcement Learning techniques: Q-Learning and Sarsa. Ultimately, the output of this work presents a benchmarking system for robotics that allows different techniques and algorithms to be compared using the same virtual conditions.

Cited by in corpus (28)

• PettingZoo: Gym for Multi-Agent Reinforcement Learning
• CHALET: Cornell House Agent Learning Environment
• ns3-gym: Extending OpenAI Gym for Networking Research
• ReLMoGen: Leveraging Motion Generation in Reinforcement Learning for Mobile Manipulation
• gym-gazebo2, a toolkit for reinforcement learning using ROS 2 and Gazebo
• TriFinger: An Open-Source Robot for Learning Dexterity
• Deep Reinforcement Learning based Local Planner for UAV Obstacle Avoidance using Demonstration Data
• Motion Planning by Reinforcement Learning for an Unmanned Aerial Vehicle in Virtual Open Space with Static Obstacles
• Flight Controller Synthesis Via Deep Reinforcement Learning
• ROS2Learn: a reinforcement learning framework for ROS 2
• OffWorld Gym: open-access physical robotics environment for real-world reinforcement learning benchmark and research
• Planning Robot Motion using Deep Visual Prediction
• Reinforcement Learning-Based Coverage Path Planning with Implicit Cellular Decomposition
• Learning to Sequence Robot Behaviors for Visual Navigation
• Sequential Learning of Movement Prediction in Dynamic Environments using LSTM Autoencoder
• Visual Exploration and Energy-aware Path Planning via Reinforcement Learning
• The Adversarial Resilience Learning Architecture for AI-based Modelling, Exploration, and Operation of Complex Cyber-Physical Systems
• Hierarchical Learning for Modular Robots
• The Chef's Hat Simulation Environment for Reinforcement-Learning-Based Agents
• Danger-aware Adaptive Composition of DRL Agents for Self-navigation
• Towards self-adaptable robots: from programming to training machines
• Sim-Env: Decoupling OpenAI Gym Environments from Simulation Models
• A Scalable and Reproducible System-on-Chip Simulation for Reinforcement Learning
• Plantbot: A New ROS-based Robot Platform for Fast Building and Developing
• Implementation of Q Learning and Deep Q Network For Controlling a Self Balancing Robot Model
• Setting Out a Software Stack Capable of Hosting a Virtual ROS-based Competition
• Composable Learning with Sparse Kernel Representations
• ASIST: Annotation-free Synthetic Instance Segmentation and Tracking by Adversarial Simulations