Uncertainty-Aware Non-Prehensile Manipulation with Mobile Manipulators under Object-Induced Occlusion

2026-02-02

AI Summary

This work addresses the challenge of limited perception in mobile manipulation tasks due to object occlusion, which often leads to collisions. The authors propose CURA-PPO, a reinforcement learning framework that uniquely integrates perception uncertainty modeling with active perception. By predicting a distribution over collision likelihoods to quantify risk and leveraging confidence maps to guide the robot in actively adjusting its viewpoint during manipulation, the method mitigates occlusion effects. Evaluated across diverse object sizes and obstacle configurations, CURA-PPO significantly improves task performance, achieving up to three times higher success rates than baseline approaches, particularly in scenarios with severe sensor occlusion.

Technology Category

• Intelligent Robots: Manipulation
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Non-prehensile manipulation using onboard sensing presents a fundamental challenge: the manipulated object occludes the sensor's field of view, creating occluded regions that can lead to collisions. We propose CURA-PPO, a reinforcement learning framework that addresses this challenge by explicitly modeling uncertainty under partial observability. By predicting collision possibility as a distribution, we extract both risk and uncertainty to guide the robot's actions. The uncertainty term encourages active perception, enabling simultaneous manipulation and information gathering to resolve occlusions. When combined with confidence maps that capture observation reliability, our approach enables safe navigation despite severe sensor occlusion. Extensive experiments across varying object sizes and obstacle configurations demonstrate that CURA-PPO achieves up to 3X higher success rates than the baselines, with learned behaviors that handle occlusions. Our method provides a practical solution for autonomous manipulation in cluttered environments using only onboard sensing.

Problem

• non-prehensile manipulation
• object-induced occlusion
• mobile manipulators
• sensor occlusion
• partial observability

Innovation

• uncertainty-aware manipulation
• non-prehensile manipulation
• partial observability
• active perception
• reinforcement learning