The iRobotCNC robot team from HUST School of Mechanical Science and Engineering won the IROS 2020 CHALLENGE OCRTOC (IROS 2020: Open Cloud Robot Table Organization Challenge) beating 59 other teams to win the 10,000 USD prize and overall championship. The iRobotCNC team was guided by Academician Han Ding and Professor Huan Zhao from State Key Laboratory of Digital Manufacturing Equipment and Technology. Its members include Jiyang Guo，Dong Liu，Hui Wang，Yecan Yinand Dongsheng Ge.
In the competition, iRobotCNC successfully solved tasks including cloud robot remote target recognition and the grasping and manipulation of objects. This was achieved by adopting algorithms including point cloud segmentation clustering, desktop fitting segmentation, semantic recognition and pose estimation based on Nearest Neighbor Distance Ratio, etc.
The competition was held online during the IROS 2020 International Conference on Intelligent Robots and Systems. It focused on table organization capability for service robots, and aimed to narrow the gap between requirements for daily-used robotic grasping and manipulation systems and the current level of state-of-the-art methods. Hosted by Alibaba AI Labs, the event attracted 59 teams from HUST, Carnegie Mellon University, Princeton University, Hong Kong University of Science and Technology, Tsinghua University and many other universities.
The IROS 2020 International Conference on Intelligent Robots and Systems was co-hosted by the IEEE Robotics and Automation Society (RAS), Robotics Society of Japan (RSJ), the IEEE Industrial Electronics Society (IES) and other organizations. The conference and IEEE International Conference on Robotics and Automation (ICRA) represent two of the largest and most influential flagship scholar conferences within the field of robotics.
Written by: Jingyi Ye
Edited by: Rong He, Andrew, Yumeng Peng
Source: School of Mechanical Science and Engineering