UNCLE-Grasp: Uncertainty-Aware Grasping of Leaf-Occluded Strawberries

2026-01-20

AI Summary

This work addresses the challenge of robotic strawberry grasping under partial occlusion caused by foliage, where conventional approaches often fail due to their reliance on a single deterministic shape estimate. The authors propose an uncertainty-aware grasping framework that generates multiple shape hypotheses via point cloud completion and quantifies geometric uncertainty using Monte Carlo Dropout. For each hypothesis, candidate grasps are generated and evaluated for feasibility using a force-closure metric. A conservative decision is then made based on a lower confidence bound criterion: high-risk grasps are deliberately rejected under severe occlusion, while high success rates are maintained when sufficient visual information is available. This approach is the first to explicitly incorporate shape completion uncertainty into grasp planning, demonstrating significant performance gains over deterministic baselines in both simulation and real-world experiments.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Robotic strawberry harvesting is challenging under partial occlusion, where leaves induce significant geometric uncertainty and make grasp decisions based on a single deterministic shape estimate unreliable. From a single partial observation, multiple incompatible 3D completions may be plausible, causing grasps that appear feasible on one completion to fail on another. We propose an uncertainty-aware grasping pipeline for partially occluded strawberries that explicitly models completion uncertainty arising from both occlusion and learned shape reconstruction. Our approach uses point cloud completion with Monte Carlo dropout to sample multiple shape hypotheses, generates candidate grasps for each completion, and evaluates grasp feasibility using physically grounded force-closure-based metrics. Rather than selecting a grasp based on a single estimate, we aggregate feasibility across completions and apply a conservative lower confidence bound (LCB) criterion to decide whether a grasp should be attempted or safely abstained. We evaluate the proposed method in simulation and on a physical robot across increasing levels of synthetic and real leaf occlusion. Results show that uncertainty-aware decision making enables reliable abstention from high-risk grasp attempts under severe occlusion while maintaining robust grasp execution when geometric confidence is sufficient, outperforming deterministic baselines in both simulated and physical robot experiments.

Problem

• occlusion
• grasping
• uncertainty
• strawberry harvesting
• shape completion

Innovation

• uncertainty-aware grasping
• point cloud completion
• Monte Carlo dropout
• force-closure
• lower confidence bound