panda-gym: Open-source goal-conditioned environments for robotic learning

2021/06/25 by Quentin Gallouédec, Nicolas Cazin, Gallouédec, Quentin +5 · 37 citations
Computer Science · #Artificial intelligence #Baseline (sea) #Computer science #FOS: Computer and information sciences #Human–computer interaction #Machine Learning (cs.LG) #Open source #Operating system #Programming language #Reinforcement Learning in Robotics #Reinforcement learning #Robot #Set (abstract data type) #Software #Software engineering #Stack (abstract data type) #cs.LG

paper · pdf · doi:10.48550/arxiv.2106.13687

published in arXiv (Cornell University) (Cornell University) · NeurIPS 2021 Workshop on Robot Learning: Self-Supervised and Lifelong Learning

openalex publication_date 2021/06/25 · arxiv created 2021/12/19 · arxiv updated 2021/12/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper presents panda-gym, a set of Reinforcement Learning (RL) environments for the Franka Emika Panda robot integrated with OpenAI Gym. Five tasks are included: reach, push, slide, pick & place and stack. They all follow a Multi-Goal RL framework, allowing to use goal-oriented RL algorithms. To foster open-research, we chose to use the open-source physics engine PyBullet. The implementation chosen for this package allows to define very easily new tasks or new robots. This paper also presents a baseline of results obtained with state-of-the-art model-free off-policy algorithms. panda-gym is open-source and freely available at https://github.com/qgallouedec/panda-gym.

Cited by

• Automating the Refinement of Reinforcement Learning Specifications
• MagBotSim: Physics-Based Simulation and Reinforcement Learning Environments for Magnetic Robotics
• Self-Supervised Multisensory Pretraining for Contact-Rich Robot Reinforcement Learning
• ROPES: Robotic Pose Estimation via Score-Based Causal Representation Learning
• U-LAG: Uncertainty-Aware, Lag-Adaptive Goal Retargeting for Robotic Manipulation
• Controlling Intent Expressiveness in Robot Motion with Diffusion Models
• Parallel Heuristic Search as Inference for Actor-Critic Reinforcement Learning Models
• Fine-Tuning Robot Policies While Maintaining User Privacy
• Robotic Manipulation via Imitation Learning: Taxonomy, Evolution, Benchmark, and Challenges
• Mutual Information Tracks Policy Coherence in Reinforcement Learning
• DEMONSTRATE: Zero-shot Language to Robotic Control via Multi-task Demonstration Learning
• Learning Task Belief Similarity with Latent Dynamics for Meta-Reinforcement Learning
• Can you see how I learn? Human observers' inferences about Reinforcement Learning agents' learning processes
• CIRO7.2: A Material Network with Circularity of -7.2 and Reinforcement-Learning-Controlled Robotic Disassembler
• Integrated Sensing, Computing, Communication, and Control for Time-Sequence-Based Semantic Communications
• GRAML: Goal Recognition As Metric Learning
• Effective Reinforcement Learning Control using Conservative Soft Actor-Critic
• CiRL: Open-Source Environments for Reinforcement Learning in Circular Economy and Net Zero
• General Dynamic Goal Recognition
• Improvements of Dark Experience Replay and Reservoir Sampling towards Better Balance between Consolidation and Plasticity
• Next-Future: Sample-Efficient Policy Learning for Robotic-Arm Tasks
• TW-CRL: Time-Weighted Contrastive Reward Learning for Efficient Inverse Reinforcement Learning

Related

• Open-Source Reinforcement Learning Environments Implemented in MuJoCo with Franka Manipulator
• Extending the OpenAI Gym for robotics: a toolkit for reinforcement learning using ROS and Gazebo
• Multi-Goal Reinforcement Learning: Challenging Robotics Environments and Request for Research
• gym-gazebo2, a toolkit for reinforcement learning using ROS 2 and Gazebo
• Scilab-RL: A software framework for efficient reinforcement learning and cognitive modeling research