MedVLA improves safety and accuracy in medical robot surgery
MedVLA: A Hierarchical Vision-Language-Action Framework for Closed-Loop Precision Medical Robot Manipulation
Robotics
Summary
Medical robots need to make smart decisions carefully to safely perform precise tasks. The authors created MedVLA, a system that combines high-level understanding of images and language with strict control over robot actions. MedVLA was tested on a robot implanting electrodes 100 times and succeeded 95% of the time, far better than other methods. This suggests using structured reasoning plus strict action rules can help medical robots work safely and reliably.
What this means in practice
- •For medical robotics teams: Develop closed-loop robotic systems that safely and accurately perform precise surgical manipulations using structured vision-language reasoning.
- •For industrial robotics engineers: Integrate hierarchical action frameworks with multimodal reasoning to improve precision and safety in complex robot tasks beyond medical use.
Authors
Junjie Xie, Chuxuan He, Angen Ye, Yujia Song, Dapeng Zhang
Abstract
Precision medical robotics demands adaptive decision-making under strict safety, interpretability, and execution constraints. Although recent Vision-Language-Action (VLA) models show strong multimodal reasoning ability, their continuous action generation paradigm is not well suited for precision medical tasks, where reliable closed-loop operation may also depend on non-action system function calls. To address this gap, we propose MedVLA, a hierarchical framework that couples high-level multimodal reasoning with low-level function-constrained execution. We further introduce a scalable multi-agent pipeline to generate skill-oriented chain-of-thought(CoT) data for structured training. Built on different multimodal large-model backbones, MedVLA consistently improves performance after fine-tuning, demonstrating the effectiveness of the proposed framework across model variants. Under identical initial conditions, we perform 100 closed-loop flexible electrode implantation trials. The results show that MedVLA achieves a 95.0\% task success rate, substantially outperforming representative VLA baselines, including OpenVLA (8\%) and $π_0$ (15\%), in accuracy, stability, and safety. These results indicate that structured reasoning with constrained function-level execution is a practical route toward deployable precision medical robotics.