【摘要】 经自然腔道手术机器人要求具有良好的灵巧性，但灵巧性的提高会产生运动学模型复杂且难以建模的问题。针对绳驱动的多段柔性连续体机械臂进行了运动学分析和实验研究。首先，以双段柔性连续体机械臂为原型，基于分段常曲率和神经网络的方法，建立了任务空间和驱动空间的正逆运动学映射模型。以任务空间变量连续体机械臂末端位置为神经网络的输入量，驱动空间变量驱动绳的拉伸量为神经网络的输出量，拟合出了连续体机械臂末端位置和驱动绳拉伸量的映射关系。最后，进行了连续体机械臂的弯曲运动和轨迹跟踪实验。实验结果表明：双段连续体机械臂弯曲角度可达180°,验证了连续体机械臂具有大角度弯曲能力；在连续体机械臂末端的圆轨迹跟踪实验中，连续体机械臂的末端在x轴、y轴、z轴方向的平均绝对位置误差为1.743、1.334和1.172 mm,验证了连续体机械臂的运动学模型的有效性。更多还原

【Abstract】 Transcatheter natural cavity surgery robots require good dexterity, but increased dexterity creates the problem that their kinematic models are complex and difficult to model. Kinematic analysis and experimental studies were carried out for rope-driven multi-segment soft continuum manipulator. First, with the prototype of a two-segment soft continuum manipulator, a forward and inverse kinematic mapping model between task space and drive space was established based on the segmented constant curvature and neural network approach. The mapping relationship between the end position of the continuum manipulator and the stretching amount of the drive rope was fitted using the task space variable end position of the continuum manipulator as the input to the neural network and the stretching amount of the drive space variable drive rope as the output to the neural network. Finally, the bending motion and trajectory tracking experiments of the continuum manipulator were conducted. The experimental results show that the bending angle of the two-segment continuum manipulator can reach 180°, which verifies that the manipulator has the ability of large-angle bending; in the circular trajectory tracking experiment at the end of the continuum manipulator, the average absolute position errors of the end of the continuum manipulator in the x-axis, y-axis and z-axis directions are 1.743 mm, 1.334 mm and 1.172 mm, which verifies the validity of the kinematic model of the continuum manipulator.更多还原