Flexure-Based Variable Stiffness Gripper for Large-Scale Grasping Force Regulation with Vision

This paper presents a vision grasping force sensing and regulation approach for a flexure-based variable stiffness gripper to handle broad-range objects from fragile/soft to rigid/heavy. The proposed vision approach can achieve high-precision grasping force sensing/regulation in large scale without any force sensor, which is realized base on the predictable and adjustable stiffness property of the gripper finger in our structure-controlled variable stiffness gripper, so that the deflection angle of the fingers can be detected from the vision to estimate the corresponding grasping force. Different with traditional vision-based approaches that require tedious manual calibration between the camera and target to retrieve spatial information, our approach incorporates the self-calibration algorithm to calibrate the vision system by employing the specifically-designed visual marks, so that the vision source can be placed with large flexibility in practical operation. Benefited from the large-ratio stiffness variation range of our gripper, the grasping force regulation can be achieved in a large scale, which makes it a universal gripper to handle broad-range objects with different material properties. Various experiments have been conducted to evaluate the force grasping performance using the proposed approach, including force regulation accuracy and grasping of extremely delicate objects (thin potation chip and plastic cup of water) to heavy industrial components