Research On The Design And Control Of A Flexible Continuum Manipulator With Gripper

Traditional robotic arms are difficult to fully exploit their performance in unstructured environments due to their articulated structure.In recent years,flexible continuum manipulators have received increasing attention and development for their ability of adapting to unstructured environments with continuous bending and deformation.One of the most praiseworthy achievements is the continuum manipulator as a major component of noninvasive surgical robots,which can achieve flexible motion in the restricted unstructured environment in human body for the surgeon to perform non-invasive surgery remotely.In addition,in the restricted unstructured environment of reality life(e.g.,disaster rescue),the continuum manipulator has advantages in environmental adaptability that are difficult for the articulated robotic arms.Therefore,the research on continuum manipulator has a very important role in expanding the scope of robotics applications and in increasing the popularity of robots in production life.Based on the requirements of operation tasks in complex environments,a flexible continuum manipulator with variable stiffness and scalability is developed.The characteristic of variable stiffness makes the manipulator have passive flexibility to adapt to complex environment under low stiffness,while it has stable load-operation capability under high stiffness.Scalability expands the workspace of the manipulator.The spring structure is used as the main framework of the manipulator to achieve scalability,and the variable stiffness characteristics are achieved through the antagonism between pneumaticdriving and cable-driving.Then,the body structure,driving unit and electrical system of the two-stage flexible continuum manipulator are designed,and a prototype of the flexible continuum manipulator with an overall expansion ratio of 1.53 is developed.Based on the assumption of constant curvature bending,the forward kinematics model of the continuum manipulator is derived,and the reachable motion space of the manipulator is obtained accordingly,which provides a basis for the parameter design of the manipulator for specific task requirements.Aiming at the problem of poor accuracy of model-based control method for continuum robot,an iterative feedback control method based on modeless inverse kinematics of the manipulator is proposed to improve the control accuracy.The model-based control method has poor motion accuracy and is sensitive to external load and interference due to the assumptions in modeling.In this paper,the inverse kinematics model of the manipulator is obtained through modeless method(deep learning network),then an iterative feedback control method based on the inverse kinematics model is proposed to achieve high-precision single point positioning and trajectory tracking control of the manipulator.The manipulator’s ability of resisting external load and interference is improve as well.Because the continuum manipulator has no controllable degree of freedom to rotate around its axis,the enveloping soft gripper with circumferential symmetric structure is designed as the end-effector of the manipulator to complete the grasping task.The circumferential symmetric structure makes the degrees of freedom rotating around the axis not necessary,and the enveloping-gripping method increases the contact area between the gripper and the object to achieve a more stable and effective grasp.The soft gripper relies on the fluid medium(liquid or gas)to obtain flexible grasp,while the fluidity of the fluid makes it have the adaptability to the shape of object.The enveloping soft gripper can adaptively grasp objects with different shapes,sizes and hardness.The flexible continuum manipulator and the enveloping soft gripper are combined to obtain the flexible continuum manipulator-hand system,then the function and performance evaluation experiments are carried out.In the open environment,taking the grasp-transferplace operation as the task,the robot can independently plan the task path and complete the operation task through tracking of the path stably based on the iterative learning control algorithm.For the restricted environment,the passive compliance control method for the manipulator was proposed,so that the manipulator can passively adapt to the restricted environment with low stiffness,then the manipulator can complete the grasp task with higher stiffness after passing the restricted environment.The experiments verify the feasibility,stability and effectiveness of the continuum manipulator-hand system in completing the tasks in open and restricted environments.