| As an important interface between robot and surrounding environment,manipulator has always been one of the key research directions in the field of robot.In recent years,soft gripper has been widely used in unstructured scenes because of its good flexibility,environmental adaptability and human-computer interaction.Aiming at the problem that the soft gripper can not establish an accurate grasp model in unstructured environment,based on the visual recognition information and the tactile perception ability of the gripper,the grasp strategy of the soft gripper for unknown objects is studied.(1)The design of the driving and sensing integrated soft gripper is carried out.Based on the characteristics of envelope grasping and fingertip grasping,a reasonable sensor layout is designed.Based on Yeoh model,the mathematical model of bending angle and inflation pressure of soft finger in free state is established and compared with the results of finite element analysis.Through the finite element analysis,the bending angle of fingers under different fingertip load and inflation pressure is obtained.Based on the idea of shape deposition manufacturing technology,the casting mold of distributed forming is designed,and the soft gripper is made.(2)This paper analyzes the basic principle of three-dimensional point cloud basic shape simplification algorithm and uses this algorithm to segment and simplify the object point cloud and fits it into the basic shape matching the grasping contour of the soft gripper.The optimal grasping size of the gripper is measured through experiments,and the most suitable basic shape is found according to the size to obtain the optimal grasping position.The combination of three-dimensional point cloud basic shape simplification algorithm and soft gripper grasping not only gives full play to the algorithm's ability to recognize and simplify the shape of unknown objects,but also makes up for the inaccuracy of the algorithm itself.(3)The information acquisition and wireless transmission system of soft gripper tactile perception is built.The shape perception and force perception of the soft gripper are used to optimize the grasping,and the feedback mechanism is established for the instability and failure of the grasping.A grasping strategy based on visual and tactile information fusion is proposed.(4)The operating system of soft gripper is built based on UR robot.Through experiments,the tactile sensing characteristics of the gripper are analyzed and the internal parameter calibration and hand-eye calibration of the depth camera are completed.Through a series of experiments on grasping unknown objects,the effectiveness and feasibility of the operation system are verified.The research content of this paper provides a new idea for the operation of the soft gripper on unknown objects and is of great significance for the practical application of the soft gripper. |
PDF Full Text Request