DoorGym: A Scalable Door Opening Environment And Baseline Agent

Abstract

In order to practically implement the door opening task, a policy ought to be robust to a wide distribution of door types and environment settings. Reinforcement Learning (RL) with Domain Randomization (DR) is a promising technique to enforce policy generalization, however, there are only a few accessible training environments that are inherently designed to train agents in domain randomized environments. We introduce DoorGym, an open-source door opening simulation framework designed to utilize domain randomization to train a stable policy. We intend for our environment to lie at the intersection of domain transfer, practical tasks, and realism. We also provide baseline Proximal Policy Optimization and Soft Actor-Critic implementations, which achieves success rates between 0% up to 95% for opening various types of doors in this environment. Moreover, the real-world transfer experiment shows the trained policy is able to work in the real world. Environment kit available here: https://github.com/PSVL/DoorGym/

Accepted to NeurIPS2019 Deep Reinforcement Learning Workshop. Full version

References in corpus (6)

• Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
• Soft Actor-Critic Algorithms and Applications
• DeepMind Control Suite
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
• Learning Contact-Rich Manipulation Skills with Guided Policy Search
• Sim2Real View Invariant Visual Servoing by Recurrent Control

Cited by in corpus (9)

• iGibson 1.0: a Simulation Environment for Interactive Tasks in Large Realistic Scenes
• ManiSkill: Generalizable Manipulation Skill Benchmark with Large-Scale Demonstrations
• VAT-Mart: Learning Visual Action Trajectory Proposals for Manipulating 3D ARTiculated Objects
• Invariant Policy Optimization: Towards Stronger Generalization in Reinforcement Learning
• Continual Model-Based Reinforcement Learning with Hypernetworks
• Lyceum: An efficient and scalable ecosystem for robot learning
• AdaAfford: Learning to Adapt Manipulation Affordance for 3D Articulated Objects via Few-shot Interactions
• DROID: Minimizing the Reality Gap using Single-Shot Human Demonstration
• Single RGB-D Camera Teleoperation for General Robotic Manipulation