Visual Grasping with Spectral Clustering and Heuristic Searching for Robot in Cluttered Environments
Wenjie Geng; Zhiqiang Cao; Yingbo Tang; Shuo Wang; Fengshui Jing
2022-12
Conference NameIEEE International Conference on Robotics and Biomimetics
Conference Date2022-12-5
Conference PlaceJinghong China
Abstract

Grasping the target object is an essential requirement for the robot to provide better services. It becomes complicated especially in cluttered environments, which still remains challenging. This paper proposes a novel grasping chain generation solution that enables the robot to grasp the target after other obstructed objects are moved in a good order. SSD is firstly adopted to acquire the information of detectable objects and then Euclidean clustering is employed to obtain the untrained objects. After that, the minimum bounding box of each object is obtained, which is then projected on the plane and represented by a smooth differentiable minimum ellipse. On this basis, an information density kernel function is designed to express the interaction between objects. By abstracting each ellipse as a node of the graph whose edge weight is calculated by this kernel function, the whole scene is described in a form of graph. To simplify the complexity of the scene graph, we use spectral clustering algorithm to classify the objects, and the taskoriented objects graph is constructed according to the objects closely related to the target one. As a result, the searching space is reduced. With space division of task-oriented objects graph, each candidate grasping chain is iteratively extended by using the heuristic searching and the best chain with the shortest length is determined. The proposed method solves the barrier caused by secondary obstruction and its effectiveness is testified by experiments.

Subject Area自动控制技术 ; 自动化技术应用 ; 人工智能其他学科
MOST Discipline Catalogue工学 ; 工学::控制科学与工程
Indexed ByEI
Language英语
Sub direction classification机器人感知与决策
planning direction of the national heavy laboratory认知决策知识体系
Paper associated data
Document Type会议论文
Identifierhttp://ir.ia.ac.cn/handle/173211/52254
Collection多模态人工智能系统全国重点实验室_智能机器人系统研究
Affiliation1.Institute of Automation Chinese Academy of Sciences
2.the School of Artificial Intelligence University of Chinese Academy of Sciences
Recommended Citation
GB/T 7714
Wenjie Geng,Zhiqiang Cao,Yingbo Tang,et al. Visual Grasping with Spectral Clustering and Heuristic Searching for Robot in Cluttered Environments[C],2022.
Files in This Item:
File Name/Size DocType Version Access License
Visual Grasping with(2186KB)会议论文 开放获取CC BY-NC-SAView
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Wenjie Geng]'s Articles
[Zhiqiang Cao]'s Articles
[Yingbo Tang]'s Articles
Baidu academic
Similar articles in Baidu academic
[Wenjie Geng]'s Articles
[Zhiqiang Cao]'s Articles
[Yingbo Tang]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Wenjie Geng]'s Articles
[Zhiqiang Cao]'s Articles
[Yingbo Tang]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: Visual Grasping with Spectral Clustering and Heuristic Searching for Robot in Cluttered Environments.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.