Vision-Based Safe Human-Robot Collaboration with Uncertainty Guarantees

2026-04-16

AI Summary

This work addresses the uncertainty inherent in human pose estimation and motion prediction for vision-driven human–robot collaboration by proposing a safety-oriented framework that integrates visual perception with conformal prediction. For the first time, conformal prediction is introduced into visual motion forecasting for collaborative robotics, combining stochastic uncertainty modeling with out-of-distribution (OOD) detection to produce high-confidence prediction intervals endowed with rigorous probabilistic guarantees. The framework is validated on real-world human motion datasets and in practical collaborative scenarios, demonstrating significant improvements in both system safety and prediction reliability.

Technology Category

Computer Vision: Motion & TrackingIntelligent Robots: Multimodal Perception & Sensor FusionHumans and AI: Human-Aware Planning and Behavior Prediction

Application Category

Responsible Web: Machine-in-the-loop, human agency and autonomySearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingUser Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

We propose a framework for vision-based human pose estimation and motion prediction that gives conformal prediction guarantees for certifiably safe human-robot collaboration. Our framework combines aleatoric uncertainty estimation with OOD detection for high probabilistic confidence. To integrate our pipeline in certifiable safety frameworks, we propose conformal prediction sets for human motion predictions with high, valid confidence. We evaluate our pipeline on recorded human motion data and a real-world human-robot collaboration setting.

Problem

• human-robot collaboration
• vision-based
• uncertainty guarantees
• conformal prediction
• safety certification

Innovation

• conformal prediction
• aleatoric uncertainty
• out-of-distribution detection
• human-robot collaboration
• vision-based pose estimation