Measuring Uncertainty in Shape Completion to Improve Grasp Quality

2025-04-22

AI Summary

Single-view point cloud shape completion for robotic grasping suffers from model uncertainty, leading to grasp failures. Method: This paper proposes a novel inference-time 3D shape completion uncertainty quantification method and—first in the literature—explicitly incorporates this uncertainty into the grasp pose quality scoring function. Uncertainty is estimated via Monte Carlo Dropout, and an uncertainty-weighted quality evaluation model is constructed. Physical experiments are conducted on a 7-DOF robotic arm. Contribution/Results: In real-world household object grasping tasks, the proposed uncertainty-aware ranking strategy significantly improves the success rate of the top-5 grasp candidates, outperforming state-of-the-art uncertainty-agnostic approaches. This work establishes a new paradigm for uncertainty-driven embodied intelligent decision-making.

Technology Category

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

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Shape completion networks have been used recently in real-world robotic experiments to complete the missing/hidden information in environments where objects are only observed in one or few instances where self-occlusions are bound to occur. Nowadays, most approaches rely on deep neural networks that handle rich 3D point cloud data that lead to more precise and realistic object geometries. However, these models still suffer from inaccuracies due to its nondeterministic/stochastic inferences which could lead to poor performance in grasping scenarios where these errors compound to unsuccessful grasps. We present an approach to calculate the uncertainty of a 3D shape completion model during inference of single view point clouds of an object on a table top. In addition, we propose an update to grasp pose algorithms quality score by introducing the uncertainty of the completed point cloud present in the grasp candidates. To test our full pipeline we perform real world grasping with a 7dof robotic arm with a 2 finger gripper on a large set of household objects and compare against previous approaches that do not measure uncertainty. Our approach ranks the grasp quality better, leading to higher grasp success rate for the rank 5 grasp candidates compared to state of the art.

Problem

• Estimating uncertainty in 3D shape completion for single-view point clouds
• Improving grasp quality scoring by incorporating completion uncertainty
• Enhancing robotic grasp success rates using uncertainty-aware algorithms

Innovation

• Calculates uncertainty in 3D shape completion
• Updates grasp quality score with uncertainty
• Improves grasp success rate significantly