Diffusion Stabilizer Policy for Automated Surgical Robot Manipulations

2025-03-03

AI Summary

Surgical robot automation faces challenges in real-world settings due to high noise and frequent failures in demonstration data, leading to poor robustness in policy learning. To address this, we propose DSP, a diffusion-based policy learning framework that introduces a novel two-stage diffusion stabilization training paradigm: (1) adaptive perturbation filtering via action prediction error estimation, and (2) incremental policy refinement. DSP enables reliable training and cross-task transfer directly from noisy or even failed demonstration trajectories—without requiring high-fidelity expert demonstrations. Evaluated across multiple surgical simulation environments, DSP significantly improves policy accuracy and robustness: performance degradation under perturbations is reduced by 42%, and generalization surpasses existing imitation learning approaches.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Natural Language Processing: Safety and Robustness
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Intelligent surgical robots have the potential to revolutionize clinical practice by enabling more precise and automated surgical procedures. However, the automation of such robot for surgical tasks remains under-explored compared to recent advancements in solving household manipulation tasks. These successes have been largely driven by (1) advanced models, such as transformers and diffusion models, and (2) large-scale data utilization. Aiming to extend these successes to the domain of surgical robotics, we propose a diffusion-based policy learning framework, called Diffusion Stabilizer Policy (DSP), which enables training with imperfect or even failed trajectories. Our approach consists of two stages: first, we train the diffusion stabilizer policy using only clean data. Then, the policy is continuously updated using a mixture of clean and perturbed data, with filtering based on the prediction error on actions. Comprehensive experiments conducted in various surgical environments demonstrate the superior performance of our method in perturbation-free settings and its robustness when handling perturbed demonstrations.

Problem

• Automating surgical robots for precise surgical tasks.
• Training with imperfect or failed surgical trajectories.
• Ensuring robustness in handling perturbed surgical demonstrations.

Innovation

• Diffusion-based policy learning framework
• Training with imperfect or failed trajectories
• Two-stage training with clean and perturbed data