【作者】 张国庆；

【导师】 李特；

【作者基本信息】 大连理工大学 ， 机械工程， 2024， 硕士

【摘要】 在高端装备制造领域的质量检测和设备维护等方面往往存在着环境复杂、路径曲折的狭小环境。这些狭小空间限制了传统工业机器人的进入,从而降低其工作效率。连续体机器人通常采用柔性材料,具有大长径比和灵活的形变能力等特点,能够较好的适用于狭小空间内的任务作业需求。然而,传统的连续体机器人由于单一的运动形式而在灵活性上存在不足,通常需要安装于大型进给平台或工业机器人的末端,导致设备体积与复杂性增加。为解决这一问题,我们提出了一种高集成度的运动复合型连续体机器人的设计思路。相比于传统的运动形式单一的连续体机器人,运动复合型连续体机器人更具应用潜力,通过集成多种运动形式,如弯曲、伸缩和扭转等,使其运动灵活性方面得到显著提升。运动复合型连续体机器人作业环境和自身结构的复杂性使其运动规划较为困难。因此,本文以一款绳-气混合驱动的可伸缩-弯曲连续体机器人为研究对象,以机器人路径跟踪精度和运动安全性为核心,展开针对狭小空间的末端跟随和局部避障运动规划研究。本文主要研究内容包括:针对狭小空间内的任务作业难题,本文设计了一款绳-气混合驱动的运动复合型连续体机器人。该机器人由可伸缩-弯曲柔性臂和驱动单元组成,并基于分段常曲率假设（PCC）和改进的D-H参数法完成运动学建模,并实现机器人多层级映射关系分析。针对狭小空间内的高精度在线作业难题,本文提出了一种可伸缩-弯曲连续体机器人的末端跟随运动规划方法。该方法在连续体机器人关节点更新模型的基础上,引入了目标路径自适应的辅助点模型,并基于BP神经网络完成辅助点模型参数的预测。针对狭小空间内的任务作业安全性难题,本文提出了一种可伸缩-弯曲连续体机器人的局部避障运动规划方法。该方法在碰撞检测的基础上,通过控制点的修正向量获得关节点的修正向量,从而对关节点的位置进行更新,实现机器人局部节段的姿态调节。基于连续体机器人实验平台,本文针对多组典型目标路径开展了一系列运动规划仿真实验和样机试验。仿真实验结果表明:本文所提出的方法具有良好的末端跟随能力和局部避障调姿态能力,满足在线操作实时性的需求,平均单步解算时间为4.41×10^-5 s,对各组圆弧路径的平均跟踪精度为7.8928 mm。样机实验结果表明:本文所设计的绳-气混合驱动的可伸缩-弯曲连续体机器人具有灵活的运动能力,并能够搭载可伸缩-弯曲连续体机器人运动规划算法,实现机器人的轨迹跟踪运动。更多还原

【Abstract】 Quality inspection and equipment maintenance in the field of equipment manufacturing are often characterized by narrow environments with complex environments and winding paths.These narrow spaces limit the access of traditional industrial robots,thus reducing their work efficiency.Continuum robots are usually made of flexible materials and characterized by large aspect ratios and flexible deformation capabilities,which are better suited to the needs of task operations in narrow spaces.However,conventional continuum robots are not flexible enough due to a single form of motion,and they usually need to be mounted on a large feeding platform or the end of an industrial robot,which leads to an increase in the size and complexity of the equipment.To solve this problem,we propose a design idea of a highly integrated motion-composite continuum robot.Compared with the traditional continuum robots with a single form of motion,the kinematic composite continuum robots have more potential for application,and their motion flexibility is significantly improved by integrating multiple forms of motion,such as bending,telescoping,and torsion.The complexity of the operating environment and the structure of the kinematic composite continuum robot makes its motion planning difficult.Therefore,this paper takes a rope-pneumatic hybrid-driven extendable-bending continuum robot as the research object,and focuses on the robot’s path tracking accuracy and motion safety to carry out the research of end-following and local obstacle-avoidance motion planning for the narrow space.The main research content of this paper includes:Aiming at the difficult task operation in the narrow space,this paper designs a rope-pneumatic hybrid-driven kinematic composite continuum robot.The robot consists of a retractable-curved flexible arm and a drive unit,and completes kinematic modeling based on the segmented constant curvature assumption（PCC）and the improved D-H parametric method,and realizes the analysis of the robot’s multi-level mapping relationship.Aiming at the difficult problem of high-precision online operation in a narrow space,this paper proposes an end-following motion planning method for a stretchable-curved continuum robot.Based on the joint point update model of the continuum robot,the method introduces an auxiliary point model for target path adaptation,and completes the prediction of auxiliary point model parameters based on BP neural network.Aiming at the safety challenges of task operations in a narrow space,this paper proposes a local obstacle avoidance motion planning method for a scalable-bending continuum robot.Based on collision detection,the method obtains the correction vectors of the joint points through the correction vectors of the control points,so as to update the positions of the joint points and realize the attitude adjustment of the robot’s local segments.In this paper,a series of motion planning simulation experiments and prototype experiments are carried out based on the continuum robot experimental platform for several groups of typical target paths.The results of the simulation experiments show that the proposed method has good end-following ability and local obstacle avoidance and posture adjustment ability,which meets the real-time demand of online operation,with an average single-step solving time of 4.41×10^-5 s,and an average tracking accuracy of 7.8928 mm for each group of arc paths.The results of the prototype experiments show that the rope-pneumatic hybrid-driven scalable and bendable continuum robot designed in this paper has flexible motion capability,and can be equipped with the motion planning algorithm of scalable and bendable continuum robot,realize the path tracking motion of the robot.更多还原