JAIST Repository >
School of Information Science >
Articles >
Journal Articles >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10119/16708

Title: A Hybrid 2-Stage Method for Robotic Planar Pushing
Authors: Gao, Ziyan
Elibol, Armagan
Chong, Nak Young
Issue Date: 2020-06
Publisher: Korea Robotics Society
Magazine name: Proceedings of the 2020 17th International Conference on Ubiquitous Robots (UR)
Abstract: Robotic manipulation has been applied to a particular setup and a limited number of known objects. In order to cope with these limitations, robots need to be capable of manipulating novel objects. In this work, we proposed a computationally efficient 2-stage framework for planar pushing, allowing a robot to push novel objects to a specified pose with a minimum number of steps. We developed three modules: Coarse Action Predictor (CAP), Forward Dynamic Estimator (FDE), and Physical Property Estimator (PPE). CAP predicts a mixture of Gaussian distribution of actions. FPE learns the causality between action and successive object state. PPE based on Recurrent Neural Network predicts the physical center of mass (PCOM) from the visual center of mass (VCOM) and robot-object interaction. Our preliminary experiments show the promising results to meet the required capability of pushing novel objects.
Rights: Ziyan Gao, Armagan Elibol, Nak Young Chong, A Hybrid 2-Stage Method for Robotic Planar Pushing, Proceedings of the 2020 17th International Conference on Ubiquitous Robots (UR), Kyoto, Japan, June 22-26, Late Breaking Results Paper, 2020. This material is posted here with permission of Korea Robotics Society (KROS).
URI: http://hdl.handle.net/10119/16708
Material Type: author
Appears in Collections:b10-1. 雑誌掲載論文 (Journal Articles)

Files in This Item:

File Description SizeFormat
C5_UR20_0191_FI.pdf545KbAdobe PDFView/Open

All items in DSpace are protected by copyright, with all rights reserved.

 


Contact : Library Information Section, Japan Advanced Institute of Science and Technology