Vision-based grasping force perception and servo control of robotic flexible gripper

Abstract

Abstract Robotic flexible grasping has important applications in industrial, agricultural, and other fields, where force perception is the key for robots to achieve stable grasping operations. The force perception capabilities of existing flexible grippers are primarily achieved through the use of force sensors or by embedding smart materials within them. However, these methods pose challenges such as complex installation and fixation of force sensors, high manufacturing requirements for flexible grippers, and alteration of the grasping contact surface characteristics. This study proposes a machine vision-based grasping force perception and servo control strategy for flexible grippers. Targeting the common fin-type flexible gripper structure, this study develops a grasping force perception algorithm for flexible grippers using finite element method (FEM). The algorithm primarily includes contact node identification and force perception during grasping. The grasping force perception strategy for the flexible gripper is developed using machine vision. A visual servo control system of flexible gripper grasping force is designed, and the flexible gripper grasping force perception experimental system is built. The force perception and servo control capabilities of the flexible gripper are experimentally validated. The results confirm that the proposed grasping force perception strategy enables contact force sensing and servo control, without relying on force sensors, when gripping objects of varying diameters and at different grasping nodes.