Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 28, 2026

Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects
07:36

Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects

Published on: November 30, 2018

Uncertainty-Calibrated Safety Gating for Vision-Language- Action Manipulation Under Domain Shift: Reliability Gains

Atef M Ghaleb1, Ali S Allahloh2, Sobhi Mejjaouli1

  • 1Department of Industrial Engineering, College of Engineering and Advanced Computing, Alfaisal University, Riyadh 11533, Saudi Arabia.

Sensors (Basel, Switzerland)
|May 27, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A Novel Technique for Quality Control of Microinjection Molding.

Micromachines·2026
Same author

An Analysis of Body Language of Patients Using Artificial Intelligence.

Healthcare (Basel, Switzerland)·2022
Same author

A Generalized Laser Simulator Algorithm for Mobile Robot Path Planning with Obstacle Avoidance.

Sensors (Basel, Switzerland)·2022
Same author

Body Language Analysis in Healthcare: An Overview.

Healthcare (Basel, Switzerland)·2022
Same author

Determining the time needed for workers to acclimatize to hypoxia.

International journal of biometeorology·2020
Same author

Autonomous Road Roundabout Detection and Navigation System for Smart Vehicles and Cities Using Laser Simulator-Fuzzy Logic Algorithms and Sensor Fusion.

Sensors (Basel, Switzerland)·2020
This summary is machine-generated.

Calibrated gating enhances Vision-Language-Action (VLA) policies for robotics, improving success rates under domain shifts by managing uncertainty. This approach balances reliability and intervention needs for safer, more robust robotic manipulation.

Area of Science:

  • Robotics and Artificial Intelligence
  • Machine Learning for Embodied Agents
  • Computer Vision and Control Systems

Background:

  • Vision-Language-Action (VLA) policies offer promise for complex, long-horizon robotic manipulation tasks.
  • Deployment challenges arise from domain shifts, necessitating reliable uncertainty estimation and runtime assurance.
  • Existing VLA systems often lack robustness when encountering novel environmental conditions.

Purpose of the Study:

  • To evaluate a model-agnostic, uncertainty-calibrated safety-gating wrapper for VLA policies under domain shifts.
  • To quantify the impact of calibrated gating on manipulation task success and failure risk estimation.
  • To characterize the trade-off between reliability and intervention frequency in VLA supervision.

Main Methods:

Keywords:
autonomous robotsdomain shiftembodied intelligencemultimodal sensingrobot manipulationrobust decision-makingruntime assurancesafety-critical autonomyuncertainty calibrationvision–language–action

More Related Videos

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss
07:12

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss

Published on: April 11, 2025

Related Experiment Videos

Last Updated: May 28, 2026

Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects
07:36

Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects

Published on: November 30, 2018

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss
07:12

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss

Published on: April 11, 2025

  • Implemented a safety-gating wrapper that estimates online failure risk and routes control.
  • Evaluated the wrapper on two long-horizon manipulation tasks (drawer, clutter sort) in NVIDIA Isaac Sim 5.0.
  • Introduced various domain shifts: lighting, texture, occlusion, sensor, and combined variations.

Main Results:

  • Calibrated gating improved mean shifted success from 57.5% to 77.2% compared to an ungated baseline.
  • Aggregate expected calibration error decreased significantly from 0.303 to 0.100.
  • Largest gains were observed under occlusion and combined shifts, with success rates reaching up to 87.8%.

Conclusions:

  • Uncertainty-calibrated gating materially improves VLA robustness under domain shift with lower intervention rates.
  • A trade-off exists between reliability and intervention, highlighting the need for efficient intervention policy mapping.
  • Findings provide controlled simulation evidence for trustworthy VLA supervision in embodied AI.