Contact-Free Grasp Stability Prediction with In-Hand Time-of-Flight Sensors

2026-05-06

AI Summary

This work proposes a non-contact method for predicting grasp stability that overcomes the limitations of traditional approaches relying on tactile sensors, which can only assess stability after grasp execution and are prone to noise. By mounting a multi-zone time-of-flight (ToF) sensor on the distal end of a gripper, the system captures real-time geometric and pose information of target objects prior to grasping. A machine learning classifier trained on over 2,500 physical grasp trials leverages this pre-grasp data to predict stability, achieving 86.0% classification accuracy on previously unseen objects. Operating at an inference rate of 15 Hz, the method enables rapid, closed-loop feedback without physical contact, significantly improving system responsiveness and mitigating object damage or operational delays caused by failed grasps—the first demonstration of high-frequency, non-contact grasp stability assessment in a closed-loop robotic system.

Abstract

Current approaches to grasp planning for robotics demonstrate high success rates, but degrade with noisy sensors and other factors. Previous works have proposed tactile-based grasp stability classifiers to detect failures, but these approaches rely on making contact and grasping the object to do so. We propose a contact-free grasp stability predictor using multi-zone time-of-flight sensors mounted in the distal links of a gripper. Our method, as it does not require grasping the object to make a prediction, significantly speeds up the stability classification process, cycling at 15 Hz. We collected over 2,500 real-world grasps across 15 objects to train a classifier. Additionally, we conducted grasp attempts over six additional unseen objects, three for validation and model selection, and three for model testing. Our approach demonstrated strong classification performance, with an accuracy of 85.5% on validation and 86.0% on test objects.

Problem

• grasp stability prediction
• contact-free sensing
• time-of-flight sensors
• robotic grasping
• tactile-based classification

Innovation

• contact-free
• grasp stability prediction
• time-of-flight sensors
• robotic grasping
• in-hand sensing