• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (11): 201-214.doi: 10.3901/JME.260359

• 机器人及机构学 • 上一篇    

扫码分享

用于微创手术的混联机器人机构设计与轨迹分层优化

张昕, 史东豪, 陈芝翰, 徐灵敏, 李秦川   

  1. 浙江理工大学机械工程学院 杭州 310018
  • 收稿日期:2025-06-17 修回日期:2025-10-28 发布日期:2026-07-29
  • 作者简介:张昕,男,2000年出生,硕士研究生。主要研究方向为并联机器人运动学和轨迹规划。E-mail:zx15110720375@163.com;史东豪,男,1995年出生,博士研究生。主要研究方向为串联机器人运动学动力学分析和遥操作控制。E-mail:1346408490@qq.com;陈芝翰,男,1999年出生,硕士研究生。主要研究方向为串联机器人运动学和视觉识别。E-mail:1303469982@qq.com;徐灵敏(通信作者),男,1993年出生,博士,教授。主要研究方向为并/混联机器人机构构型综合、力学建模及性能设计、装备研发及应用。发表论文30余篇。E-mail:xulm@zstu.edu.cn;李秦川,男,1975年出生,博士,教授。主要研究方向为并联机器人机构学和应用技术。发表论文100余篇。E-mail:lqchuan@zstu.edu.cn
  • 基金资助:
    浙江省自然科学基金重大项目(LD25E050001)、中国科协青年人才托举工程(2023QNRC001)、浙江省自然科学基金探索项目(LMS25E050002,LZ26E050005)、浙江理工大学科研启动基金(24242189-Y)、浙江省教育厅一般科研项目(Y202456192)和浙江理工大学科研业务费专项资金(25242161-Y)资助项目。

Mechanism Design and Trajectory Layering Optimization of a Hybrid Robot for Minimally Invasive Surgery

ZHANG Xin, SHI Donghao, CHEN Zhihan, XU Lingmin, LI Qinchuan   

  1. School of Mechanical Engineering, Zhejiang Sci-Tech University, Hangzhou 310018
  • Received:2025-06-17 Revised:2025-10-28 Published:2026-07-29

摘要: 传统微创手术机器人末端采用串联结构,普遍存在刚度弱、依赖算法实现的远中心约束存在安全隐患、难以协调机构性能与轨迹特征等问题。提出一种新型混联微创手术机器人系统,其由Franka串联机械臂与自主设计的远中心并联机构构成。针对远中心并联机构,基于螺旋理论分析其自由度特性,通过闭环矢量法与几何约束关系推导其运动学逆解,并进一步结合螺旋理论评估其在工作空间内的运动/力传递性能。为优化手术机器人性能,提出一种基于串/并联结构的分层优化框架:首先通过动态运动基元示教学习从人工示范轨迹中提取轨迹特征,生成初始参考轨迹;随后以并联机构的传递性能分布图谱为指导,在保证轨迹特征的前提下,将轨迹重新规划至并联机构高性能区域,并针对不同手术需求采用不同的优化函数进行轨迹微调;最后通过对Franka串联机械臂逆运动学与静力学的综合评价,实现其末端姿态的优化筛选。实验验证阶段搭建腹部微创手术模拟平台,在仿生假人上完成小创口轨迹操作测试。本研究验证了所设计混联手术机器人在微创手术应用中的可行性,并通过串/并联机构分层优化实现了手术轨迹的泛化学习与性能优化。

关键词: 手术机器人, 远中心并联机构, 性能评估, 示教学习, 轨迹规划

Abstract: Traditional minimally invasive surgical robots employ a serial structure at the end, which generally suffer from weak stiffness, safety hazards due to potential failure of algorithm-implemented remote center of motion (RCM) constraints, and difficulty in coordinating mechanism performance with trajectory characteristics. A novel hybrid minimally invasive surgical robot system is proposed, consisting of a Franka serial manipulator and an autonomously designed distal center parallel mechanism. Regarding the RCM parallel mechanism, the degrees of freedom of the parallel RCM mechanism are analyzed using the screw theory, and the analytical kinematic inverse solution is derived through the closed-loop vector method and geometric constraint relationship. Furthermore, a motion/force transmission performance atlas within the workspace is constructed to quantify the transmission performance of the mechanism. To optimize the performance of surgical robots, a hierarchical optimization framework based on serial/parallel structure is proposed. Firstly, trajectory features are extracted from manually demonstrated trajectories through dynamic motion primitive teaching learning, and initial optimization paths are generated. Subsequently, guided by the transmission performance atlas, the reference path is replanned to the high-performance region of the parallel mechanism while ensuring its characteristics. Depending on the different surgical requirements, different optimization functions are employed for fine-tuning. Finally, the posture of the serial robotic arm is optimized and selected through inverse kinematics analysis and comprehensive evaluation of static performance. During the experimental verification phase, a minimally invasive abdominal surgery simulation platform is built, and small incision trajectory operation tests are successfully conducted on a biomimetic prosthesis. This study validates the feasibility of the proposed hybrid surgical robot for minimally invasive surgery, and achieves generalized learning and performance optimization of surgical trajectories through hierarchical optimization of serial/parallel mechanisms.

Key words: surgical robot, RCM parallel mechanism, performance evaluation, teaching demonstration learning, trajectory planning

中图分类号: