Algorithm analysis of finger pressing action for robotic arm with dexterous hand

In order to understand the algorithm implementation of finger pressing action for robotic arms with dexterous hands, an analysis and research on the algorithm implementation of finger pressing action for robotic arms with dexterous hands is proposed. This paper designs an intelligent data acquisition system platform based on a robotic arm dexterous hand. The platform consists of Kinect 2.0 camera, BH8-282 three finger dexterous hand, UR5 six degree of freedom robotic arm and other devices. Through automatic recognition and positioning of the target object, the autonomous motion planning path completes the grasping action of the target object, and obtains visual and tactile information of the grasped target object. Experiments have shown that the platform can effectively grasp target objects without supervision and achieve fully automated collection of visual and tactile data. During the experimental process, data analysis was conducted on 3590 groups of target objects. The proportion of objects that were not captured was 16.42%, the proportion of objects that were firmly grasped was 43.92%, and the proportion of objects that were not firmly grasped was 42.69%. Overall, the experimental effect was good.