Grasp the Graph (GtG) 2.0: Ensemble of GNNs for High-Precision Grasp Pose Detection in Clutter

2025-05-05

AI Summary

To address the challenge of 7-degree-of-freedom (7-DoF) grasp pose detection in cluttered real-world scenes—where point cloud noise, occlusion, and complex object geometry severely degrade performance—this paper proposes a lightweight hypothesis-and-verification framework. Our method introduces a novel dual-domain graph representation that jointly models both surface points and interior-sampled points of target objects. Furthermore, we design a multi-graph neural network (GNN) collaborative evaluation mechanism to jointly score and filter 7-DoF grasp candidates. Unlike existing approaches such as GraspNet—which assume complete, noise-free point clouds and support only 4-DoF grasps—our framework operates robustly under realistic sensing conditions. On the GraspNet-1Billion benchmark, it achieves a 35% average precision improvement and ranks among the top three methods. Physical experiments on a Delta robot demonstrate a 91% grasp success rate and 100% scene-clearing capability in highly cluttered environments.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Grasp pose detection in cluttered, real-world environments remains a significant challenge due to noisy and incomplete sensory data combined with complex object geometries. This paper introduces Grasp the Graph 2.0 (GtG 2.0) method, a lightweight yet highly effective hypothesis-and-test robotics grasping framework which leverages an ensemble of Graph Neural Networks for efficient geometric reasoning from point cloud data. Building on the success of GtG 1.0, which demonstrated the potential of Graph Neural Networks for grasp detection but was limited by assumptions of complete, noise-free point clouds and 4-Dof grasping, GtG 2.0 employs a conventional Grasp Pose Generator to efficiently produce 7-Dof grasp candidates. Candidates are assessed with an ensemble Graph Neural Network model which includes points within the gripper jaws (inside points) and surrounding contextual points (outside points). This improved representation boosts grasp detection performance over previous methods using the same generator. GtG 2.0 shows up to a 35% improvement in Average Precision on the GraspNet-1Billion benchmark compared to hypothesis-and-test and Graph Neural Network-based methods, ranking it among the top three frameworks. Experiments with a 3-Dof Delta Parallel robot and Kinect-v1 camera show a success rate of 91% and a clutter completion rate of 100%, demonstrating its flexibility and reliability.

Problem

• Detecting grasp poses in cluttered environments with noisy data
• Improving geometric reasoning using Graph Neural Networks
• Enhancing grasp detection performance for 7-Dof grasping

Innovation

• Ensemble of Graph Neural Networks for grasp detection
• 7-Dof grasp candidates from point cloud data
• Improved representation with inside and outside points