Design Of Shape Adaptive Compliant Grasper And Research Of Grasping Force Sensing Method

With the expansion of robot application scenarios,more and more robots have to perform tasks in unknow environments,which requires robots to adapt to the unknown environments and realize complex grasping or manipulation tasks.As one of the most important parts of the robot,it's essential for graspers to have abilities of grasping objects with arbitrary shapes as well as sensing grasping forces.However,most graspers have some problems,such as poor adaptive ability,low force sensing accuracy,complex control and high cost,which are difficult to meet the requirements of grasping tasks in unstructured environments.In order to solve these shortcomings,a new compliant grasper was designed,a prototype was developed and a series of related experimental research was carried out.According to the demand of adaptive grasping,a compliant grasper b ased on bionics was designed,which could grasp objects of arbitrary shape adaptively.The finger was designed based on the concept of bionic fin,which has unique deformation characteristics and can form an adaptive envelope to the object when it is under force.The fingers of the grasper were processed with flexible material,which gives the fingers the soft character and enables them to grasp without damaging the object.There is only one stepper motor in the compliant grasper,it drives two compliant fingers to complete grasping action through the device of ball screw.At the same time,using the self-developed small control board and micro-stepper motor driver to control the compliant grasper,which makes the structure very compact.The design scheme is simple and reliable,which can meet the adaptive grasping of any shape object,and the cost is low.The force and shape model of compliant grasper was established,which includes "force-shape" model and "shape-force" model.The "force-shape" model was established based on the finite element theory and the structural characteristics of compliant fingers,and it has high accuracy.Generating a large number of data about force and shape though the “force-shape” model to train the BP neural network.In this way,the "force-shape" model was established and it could achieve the goal of sensing force according to the shape of the compliant finger.The force and deformation model provides a theoretical basis for the force sensing of compliant finger.The prototype integration and experimental platform were completed,and on this basis,the calibration experiment of young's modulus of compliant finger,the force sensing experiment of single finger,the grasping force sensing experiment of compliant grasper and the grasping control experiment of compliant grasper were carried out,and the proposed adaptive grasping and force sensing method were verified.The experimental results showed that the compliant grasper can grasp objects of arbitrary shapes with high accuracy of force sensing and can control the grasping force,so it can meet the requirements of complex operation tasks.