Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Robust Adaptive Safe Robotic Grasping with Tactile Sensing

Yitaek Kim, Jeeseop Kim, Albert H. Li, Aaron D. Ames, Christoffer Sloth

TL;DR

This work tackles safe robotic grasping under contact dynamics uncertainty by introducing a CBF-based framework that enforces both contact-force ranges and force-closure constraints. It integrates multiple safety filters (CBF, RaCBF, RCBF, DOBCBF) with tactile-based contact point/force estimation and a fingertip force controller, validated in simulation and on a Shadow Hand with fragile objects. The results show that DOBCBF delivers formal safety guarantees with the least conservatism, outperforming baseline CBF approaches and enabling safe manipulation in real-world tasks. The framework advances practical safe grasping by combining formal safety analysis, adaptive disturbance handling, and tactile sensing for robust, real-time control.

Abstract

Robotic grasping requires safe force interaction to prevent a grasped object from being damaged or slipping out of the hand. In this vein, this paper proposes an integrated framework for grasping with formal safety guarantees based on Control Barrier Functions. We first design contact force and force closure constraints, which are enforced by a safety filter to accomplish safe grasping with finger force control. For sensory feedback, we develop a technique to estimate contact point, force, and torque from tactile sensors at each finger. We verify the framework with various safety filters in a numerical simulation under a two-finger grasping scenario. We then experimentally validate the framework by grasping multiple objects, including fragile lab glassware, in a real robotic setup, showing that safe grasping can be successfully achieved in the real world. We evaluate the performance of each safety filter in the context of safety violation and conservatism, and find that disturbance observer-based control barrier functions provide superior performance for safety guarantees with minimum conservatism.

Robust Adaptive Safe Robotic Grasping with Tactile Sensing

TL;DR

This work tackles safe robotic grasping under contact dynamics uncertainty by introducing a CBF-based framework that enforces both contact-force ranges and force-closure constraints. It integrates multiple safety filters (CBF, RaCBF, RCBF, DOBCBF) with tactile-based contact point/force estimation and a fingertip force controller, validated in simulation and on a Shadow Hand with fragile objects. The results show that DOBCBF delivers formal safety guarantees with the least conservatism, outperforming baseline CBF approaches and enabling safe manipulation in real-world tasks. The framework advances practical safe grasping by combining formal safety analysis, adaptive disturbance handling, and tactile sensing for robust, real-time control.

Abstract

Robotic grasping requires safe force interaction to prevent a grasped object from being damaged or slipping out of the hand. In this vein, this paper proposes an integrated framework for grasping with formal safety guarantees based on Control Barrier Functions. We first design contact force and force closure constraints, which are enforced by a safety filter to accomplish safe grasping with finger force control. For sensory feedback, we develop a technique to estimate contact point, force, and torque from tactile sensors at each finger. We verify the framework with various safety filters in a numerical simulation under a two-finger grasping scenario. We then experimentally validate the framework by grasping multiple objects, including fragile lab glassware, in a real robotic setup, showing that safe grasping can be successfully achieved in the real world. We evaluate the performance of each safety filter in the context of safety violation and conservatism, and find that disturbance observer-based control barrier functions provide superior performance for safety guarantees with minimum conservatism.

Paper Structure

This paper contains 21 sections, 2 theorems, 21 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Contributions
  3. Problem Formulation
  4. Preliminaries
  5. Contact Dynamics Models
  6. Robust Adaptive Control Barrier Functions
  7. Robust Control Barrier Functions
  8. Nonlinear Disturbance Observer (NDOB)
  9. The Proposed Framework
  10. Safety-Critical Grasping
  11. Contact Force constraint
  12. Force Closure
  13. Safety Filters for Safe Grasping
  14. Fingertip Contact Force Controller
  15. Estimating Contact Points, Forces, and Torques
  16. ...and 6 more sections

Key Result

Theorem 1

The function $h_r$ is a Robust adaptive Control Barrier Function (RaCBF) for prob:sys_linearized_model if there exists an extended class $\mathcal{K}_\infty$ function $\alpha(\cdot)$ such that with $\lambda(x,\bm{\theta}\xspace) \triangleq {\bm{\theta}\xspace} - \Gamma(\frac{\partial h_r}{\partial \bm{\theta}\xspace}(x,\bm{\theta}\xspace))^{\top}$, for any $\bm{\hat{\theta}}\xspace \in \Theta\xsp

Figures (6)

  • Figure 1: Safe robotic grasping for fragile lab glassware.
  • Figure 2: Overview of the proposed safe grasping framework. The framework consists of three main components: safety filters, estimation of contact information, and fingertip force controller. Safety filters include CBF, RaCBF, RCBF, and DOBCBF with a disturbance observer. Tactile sensor data from the hand is mapped to the actual force, and the contact force/torque and point on a fingertip are estimated to be used in the controllers. The fingertip force controller is designed to track safe control input to the hand.
  • Figure 3: The performance of contact point estimation based on tactile sensors
  • Figure 4: (a) and (b) show the simulation and real experiment setups, respectively.
  • Figure 5: (a) shows contact normal force on the contact surface, and (b) presents the control barrier function, $h_1$ which regulates the contact normal force from \ref{['force_min_lim']}. (c) shows the contact force is in the friction cone from \ref{['force_closure_const']} to achieve force closure. Lastly, (d) plots the disturbance estimation error bound and the actual error.
  • ...and 1 more figures

Theorems & Definitions (2)

  • Theorem 1: lopez2020robust
  • Theorem 2: JANKOVIC2018359