DoorGym: A Scalable Door Opening Environment And Baseline Agent

TL;DR

DoorGym is an open-source simulation environment that uses domain randomization to train robust door-opening policies, demonstrating successful transfer from simulation to real-world applications.

Contribution

We introduce DoorGym, a scalable simulation framework for training door-opening policies with domain randomization, and provide baseline RL implementations with real-world transfer capability.

Findings

Baseline policies achieve up to 95% success rate in simulation.

Trained policies successfully transfer to real-world door opening.

DoorGym enables robust policy training across diverse door types.

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/