CoAS-Net: Context-Aware Suction Network with a Large-Scale Domain Randomized Synthetic Dataset

• Published in: IEEE robotics and automation letters Vol. 9; no. 1; pp. 1 - 8
• Main Authors: Son, Yeong Gwang, Bui, Tat Hieu, Hong, Juyong, Kim, Yong Hyeon, Moon, Seung Jae, Kim, Chun Soo, Rhee, Issac, Kang, Hansol, Choi, Hyouk Ryeol
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Annotations; Collision avoidance; Computer Vision for Automation; Context; Data Sets for Robotic Vision; Datasets; Deep Learning in Grasping and Manipulation; Grasping; Grasping (robotics); Object oriented modeling; Outliers (statistics); Picking; Pipelines; Robotics; Robots; Seals; Springs; Suction; Suction Grasping; Synthetic data
• DOI: 10.1109/LRA.2023.3337692
• ISSN: 2377-3766

Robotic grasping is one of the essential skills in robotics. From industrial to housework, robots are required to handle objects, enabling them to interact with their surroundings. Among the various tasks in robotic grasping, bin-picking is considered one of the most challenging because of the cluttered bin filled with objects. Also, for the next-level automation, they need to handle unseen objects and discriminate target objects and outliers. This paper proposes a novel dataset generation pipeline for suction-grasping in bin-picking tasks. This pipeline consists of a series of methods that progressively transit from a single object evaluation to an entire scene evaluation and lower the dimension of the labels to the image space. We trained a suction prediction FCN (Fully Convolution Network) with our dataset generated from the pipeline and conducted bin-picking experiments. Our large-scale collision-free annotation enables the network to understand the context of a bin-picking task, where collisions between the gripper and the bin or object are a concern, and distinguishing the background is crucial. The results show that our solution excels the existing methods, and the network demonstrates its context-aware grasp on objects with loosely defined RoI (Region of Interest). Our dataset and the grasp detection model are available at https://github.com/SonYeongGwang/CoAS-Net.git.