Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Synchronized Online Friction Estimation and Adaptive Grasp Control for Robust Gentle Grasp

Zhenwei Niu, Xiaoyi Chen, Jiayu Hu, Zhaoyang Liu, Xiaozu Ju

TL;DR

This work tackles robust gentle grasping under uncertain and time-varying friction. It introduces a particle-filter-based online estimator for the friction coefficient $\mu_t$ that uses tangential contact measurements from Vision-Based Tactile Sensors and models friction as a stochastic state with random-walk dynamics. A tightly coupled estimation-control architecture integrates this estimator with a reactive grasp controller, employing a proportional law $\Delta x_g(t) = K_p( cf_{target} - cf_t)$ and continuously updating the contact coefficient $cf_t$ via tactile feedback. Experiments on a UR5e with Tac3D demonstrate rapid convergence of $\mu_t$, robust grasp stability during changing friction/weight and motion, highlighting the practical impact for unstructured manipulation.

Abstract

We introduce a unified framework for gentle robotic grasping that synergistically couples real-time friction estimation with adaptive grasp control. We propose a new particle filter-based method for real-time estimation of the friction coefficient using vision-based tactile sensors. This estimate is seamlessly integrated into a reactive controller that dynamically modulates grasp force to maintain a stable grip. The two processes operate synchronously in a closed-loop: the controller uses the current best estimate to adjust the force, while new tactile feedback from this action continuously refines the estimation. This creates a highly responsive and robust sensorimotor cycle. The reliability and efficiency of the complete framework are validated through extensive robotic experiments.

Synchronized Online Friction Estimation and Adaptive Grasp Control for Robust Gentle Grasp

TL;DR

This work tackles robust gentle grasping under uncertain and time-varying friction. It introduces a particle-filter-based online estimator for the friction coefficient that uses tangential contact measurements from Vision-Based Tactile Sensors and models friction as a stochastic state with random-walk dynamics. A tightly coupled estimation-control architecture integrates this estimator with a reactive grasp controller, employing a proportional law and continuously updating the contact coefficient via tactile feedback. Experiments on a UR5e with Tac3D demonstrate rapid convergence of , robust grasp stability during changing friction/weight and motion, highlighting the practical impact for unstructured manipulation.

Abstract

We introduce a unified framework for gentle robotic grasping that synergistically couples real-time friction estimation with adaptive grasp control. We propose a new particle filter-based method for real-time estimation of the friction coefficient using vision-based tactile sensors. This estimate is seamlessly integrated into a reactive controller that dynamically modulates grasp force to maintain a stable grip. The two processes operate synchronously in a closed-loop: the controller uses the current best estimate to adjust the force, while new tactile feedback from this action continuously refines the estimation. This creates a highly responsive and robust sensorimotor cycle. The reliability and efficiency of the complete framework are validated through extensive robotic experiments.
Paper Structure (9 sections, 20 equations, 5 figures, 2 algorithms)

This paper contains 9 sections, 20 equations, 5 figures, 2 algorithms.

Table of Contents

  1. INTRODUCTION
  2. Contact Model Based on VBTS
  3. Synchronized Online Friction Estimation and Adaptive Grasp Control
  4. Particle Filter for Friction Coefficient Estimation
  5. Synchronized Online Friction Estimation and Adaptive Grasp Control
  6. Experiments
  7. Stable Grasp for Objects with Unstable Friction and Changing Weight
  8. Grasp Objects Stably While Robot Moving
  9. CONCLUSIONS

Figures (5)

  • Figure 1: The proposed synchronized architecture, where the particle filter-based estimator is synergistically integrated with a reactive grasp controller, while new tactile feedback from grasp continuously refines the estimation. This allows a close loop for online friction estimation and continuous force modulation, achieving real-time adaptability and grasp stability.
  • Figure 2: A grasping experiment is conducted using a plastic cup with a moistened surface, while water is incrementally added during the manipulation task.
  • Figure 3: The grasp force and friction estimation results for grasping a plastic cup with unstable friction and changing weight.
  • Figure 4: Grasp objects stably while robot moving. Robot picks up a soft toy which has high compliance. Then robot grasp stably while doing acceleration and deceleration movement. The end-effector reached a maximum speed of 3.0 $m/s$ and a maximum acceleration of 3.0 $m/s^{2}$.
  • Figure 5: The recording of grasp forces and friction estimation during grasp while robot moving. The forces are adjusted based on dynamic movement conditions to keep stable grasp while also keeps small enough grasp forces.