ORBIT-Surgical: An Open-Simulation Framework for Learning Surgical Augmented Dexterity

TL;DR

ORBIT-Surgical introduces a fast, accurate, and photorealistic physics-based simulation framework for surgical robots, enabling training, benchmarking, and sim-to-real transfer of learned surgical policies.

Contribution

It provides a novel open-simulation environment with benchmark tasks and supports reinforcement and imitation learning for surgical robots.

Findings

Successful sim-to-real transfer of learned policies

Provides 14 benchmark surgical tasks

Enables realistic synthetic data generation

Abstract

Physics-based simulations have accelerated progress in robot learning for driving, manipulation, and locomotion. Yet, a fast, accurate, and robust surgical simulation environment remains a challenge. In this paper, we present ORBIT-Surgical, a physics-based surgical robot simulation framework with photorealistic rendering in NVIDIA Omniverse. We provide 14 benchmark surgical tasks for the da Vinci Research Kit (dVRK) and Smart Tissue Autonomous Robot (STAR) which represent common subtasks in surgical training. ORBIT-Surgical leverages GPU parallelization to train reinforcement learning and imitation learning algorithms to facilitate study of robot learning to augment human surgical skills. ORBIT-Surgical also facilitates realistic synthetic data generation for active perception tasks. We demonstrate ORBIT-Surgical sim-to-real transfer of learned policies onto a physical dVRK robot. Project website: orbit-surgical.github.io