• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (17): 179-188.doi: 10.3901/JME.260610

• 特邀专栏:人形机器人与具身智能 • 上一篇    

扫码分享

面向腹腔镜手术的人形机器人力反馈遥操作:盲操作可行性及感知精度建模研究

牛孜睿1,2, 金天浩1,2, 魏逸凡1,2, 崔馨天1,2, 乔宇1,2, 郑随兵3, 史涛坪4,5, 王党校1,2   

  1. 1. 北京航空航天大学虚拟现实技术与系统全国重点实验室 北京 100191;
    2. 北京航空航天大学机械工程及自动化学院 北京 100191;
    3. 睿尔曼智能科技(北京)有限公司 北京 100144;
    4. 中国人民解放军总医院海南医院 三亚 572010;
    5. 南方医科大学珠江医院 广州 510280
  • 收稿日期:2025-11-10 修回日期:2026-05-10 发布日期:2026-11-20
  • 作者简介:牛孜睿,男,2000年出生,博士研究生。主要研究方向为机器人遥操作技术。E-mail:Niux2_MooX@buaa.edu.cn;王党校(通信作者),男,1976年出生,博士,教授,博士研究生导师。主要研究方向为机器触觉、虚拟现实技术、智能机器人。E-mail:hapticwang@buaa.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(62373021,82373233)。

Haptic Feedback Teleoperation of Humanoid Robots for Laparoscopic Surgery: A Study on the Feasibility and Perception Accuracy Modeling of Blind Manipulation

NIU Zirui1,2, JIN Tianhao1,2, WEI Yifan1,2, CUI Xintian1,2, QIAO Yu1,2, ZHENG Suibing3, SHI Taoping4,5, WANG Dangxiao1,2   

  1. 1. State Key Laboratory of Virtual Reality Technology and Systems, Beihang University, Beijing 100191;
    2. School of Mechanical Engineering and Automation, Beihang University, Beijing 100191;
    3. RealMan Intelligent Technology (Beijing) Co., Ltd., Beijing 100144;
    4. Hainan Hospital of the General Hospital, Chinese PLA General Hospital, Sanya 572010;
    5. ZhuJiang Hospital, Southern Medical University, Guangzhou 510280
  • Received:2025-11-10 Revised:2026-05-10 Published:2026-11-20

摘要: 腹腔镜手术作为一种重要的微创技术,其手术过程依赖于助手对腹腔镜的实时操作。这一过程单调、易疲劳,消耗了额外的医疗资源。利用遥操作技术采集人类演示数据,训练人形机器人进行扶镜操作有助于节约医疗资源。然而,现有用于数据采集的遥操作的系统普遍缺乏力反馈功能,导致操作者难以感知从端与组织的交互力,严重制约了训练数据质量与数据采集效率。针对这一问题,设计了一套集成力反馈功能的人形机器人腹腔镜遥操作数据采集系统,深入探讨了其在视觉受限下的“力反馈盲操作”模式的可行性与感知精度。通过建立基于力触觉的盲操作感知模型,理论分析了时延、操作者控制策略等参数对环境估计误差及系统稳定性的影响机理。实验结果表明,与纯视觉操作相比,力反馈介入能使平均接触力降低约30%,最大接触力降低约35%;即使在纯力觉盲操作模式下,其平均接触力也显著低于纯视觉操作,证实了盲操作模式在安全交互方面的可行性。研究结果证明,力反馈在显著提升操作安全性(降低交互力)与盲操作可行性方面的关键作用,为人形机器人学习安全、复杂的手术技能提供了高质量的数据采集与演示解决方案。

关键词: 腹腔镜手术, 手术机器人, 遥操作, 力触觉感知, 盲操作

Abstract: Laparoscopic surgery, as an important minimally invasive technique, relies on the assistant’s real-time operation of the laparoscope. This process is monotonous, fatigue-prone, and consumes additional medical resources. Using teleoperation technology to collect human demonstration data to train humanoid robots for camera-holding operation helps save medical resources. However, existing teleoperation systems for data collection generally lack force feedback function, making it difficult for the operator to perceive the interaction force between the slave end and the tissue, which severely restricts the quality of training data and data collection efficiency. To address this problem, a humanoid robot laparoscopic teleoperation data collection system integrated with force feedback function is designed, and the feasibility and perception accuracy of its "force feedback blind operation" mode under vision-limited conditions are deeply discussed. By establishing a blind operation perception model based on force haptics, the influence mechanism of parameters such as time delay and operator control strategy on environmental estimation error and system stability is theoretically analyzed. Experimental results show that compared with vision-only operation, the intervention of force feedback can reduce the average contact force by about 30% and the maximum contact force by about 35%; even in the pure haptic blind operation mode, its average contact force is significantly lower than that of vision-only operation, confirming the feasibility of this mode in terms of safe interaction. The critical role of force feedback in significantly enhancing operational safety (by reducing interaction forces) and enabling feasible blind manipulation is demonstrated, and a high-quality data collection and demonstration solution for humanoid robots to learn safe and complex surgical skills is provided.

Key words: laparoscopic surgery, surgical robot, teleoperation, haptic perception, blind operation

中图分类号: