Enhancing Regrasping Efficiency Using Prior Grasping Perceptions with Soft Fingertips

TL;DR

This work tackles the challenge of efficient regrasping of unknown objects by leveraging perception from the first grasp to predict the required wrench for subsequent grasps. It introduces a predictive framework that accounts for soft fingertip contact and object asymmetry, using gravity measurements to estimate the center of gravity and to compute the minimum grasping force $W_{req}=[F_{req},M_{req}]$. The method handles both two-finger and one-finger slip states through a soft-contact limit surface model, enabling rapid calculation of the necessary grasping forces without relying solely on slow real-time feedback. Experimental results show accurate force prediction (errors within ~0.18 N), broad adaptability across objects, and substantial efficiency gains when integrating prediction into the control loop, reducing grasping time and increasing success rates in regrasp scenarios.

Abstract

Grasping the same object in different postures is often necessary, especially when handling tools or stacked items. Due to unknown object properties and changes in grasping posture, the required grasping force is uncertain and variable. Traditional methods rely on real-time feedback to control the grasping force cautiously, aiming to prevent slipping or damage. However, they overlook reusable information from the initial grasp, treating subsequent regrasping attempts as if they were the first, which significantly reduces efficiency. To improve this, we propose a method that utilizes perception from prior grasping attempts to predict the required grasping force, even with changes in position. We also introduce a calculation method that accounts for fingertip softness and object asymmetry. Theoretical analyses demonstrate the feasibility of predicting grasping forces across various postures after a single grasp. Experimental verifications attest to the accuracy and adaptability of our prediction method. Furthermore, results show that incorporating the predicted grasping force into feedback-based approaches significantly enhances grasping efficiency across a range of everyday objects.

Figures (20)

• Figure 1: During the first successful grasping, the contact information is continuously and carefully monitored and the grasping force is adjusted accordingly. After slowly lifting the unknown object, its center of gravity is measured. In regrasping attempts, the required grasping force for different positions can be directly predicted before lifting the object. This approach eliminates the need for cautious lifting with continuous feedback, thereby improving grasping efficiency.
• Figure 2: Our framework designed to determine the required grasping force in both the first successful grasping and the regrasping attempts. The required wrench can be derived from current perception during the first successful grasping or from prior perception during subsequent attempts. The required grasping force is then calculated using the required wrench, the friction coefficient, and the normal vector of the contact surface.
• Figure 3: Relationship between grasping force and time $t$, divided into three steps.
• Figure 4: Force analysis depicting contact force and other external forces acting on the object.
• Figure 5: Force analysis depicting the grasping force and the force and moment required by the object.