Robot autonomous manipulation based on visual reasoning and learning

August 1, 2022, Harbin

Professor Xuguang Lan

Abstract

This talk briefly introduces the progress and challenges of robots in terms of intelligence, especially the difficulty of robots understanding, learning and manipulation in collaboration scenarios. Aiming at the difficult problems, a manipulation method based on visually physical reasoning is proposed for partially observable scenes. The interaction between learning and planning (POMDP) enables the robot to perform visual reasoning on dynamic unstructured scenes and autonomously complete tasks for specific objects. In order to further expand the learning effect on few-shot, we studied the zero-shot learning method, by constructing the orthogonal latent space and orthogonal quantization method to realize the transfer learning of the model method. Finally, we will also introduce the robot learning using hindsight trust region policy optimization on spares reward scenarios, and the multi-robot autonomous cooperation method and related applications.

Biography

Xuguang Lan received Ph.D. degree in Pattern Recognition and Intelligent System from Xi’an Jiaotong University in 2005. Currently, he is a professor at Institute of Artificial Intelligence and Robotics, and vice dean of the college of Artificial Intelligence of in Xi’an Jiaotong University.

His main research areas include computer vision, robot learning, and human-robot collaboration. He is the director of the “Coexisting-Cooperative-Cognitive Robot, Tri-Co Robot” Committee of the Automation Society, a member of a council and deputy secretary-general of the Chinese Society of Cognitive Science, etc. He is a member of the editorial board of the international journal IET “Cognitive Computation and Systems”, “Pattern Recognition and Artificial Intelligence”, and Robotics, and IEEE CYBER 2019/ICIRA2021 Program Co-Chair. He has published more than 100 papers in journals and conferences such as IEEE Transactions and CVPR/RSS, and has obtained more than 10 national invention patents. In recent years, he has presided over more than 10 scientific research projects such as the Distinguished Young Scientist/key program of National Natural Science Foundation of China, the National Science and Technology Major Project, etc. He is a senior member of IEEE.

%d 博主赞过: