Fangyi Zhang

Fangyi Zhang

Researcher in Robotics and Artificial Intelligence

QUT Centre for Robotics

Biography

Dr. Fangyi Zhang is currently a research fellow in the QUT Centre for Robotics. He was also a former PhD student in the Australian Centre for Robotic Vision (ACRV) at QUT node, supervised by Prof. Peter Corke, Dr. Jürgen Leitner, Prof. Michael Milford and Dr. Ben Upcroft. His PhD research was focused on Deep Reinforcement Learning and Transfer Learning for Robotic Reaching, and obtained the degree in 2018 with the thesis entitled “Learning Real-world Visuo-motor Policies from Simulation”. After that, he joined Alibaba DAMO Academy as a Research Scientist, doing research and development work on drone applications and data mining. His current research interests include robot learning, robotic vision, robotic manipulation, and autonomous systems.

Prior to his PhD, Fangyi obtained his B.Eng. degree in Automation from the East China Jiaotong University in 2010, followed with three years' work experience on R&D of locomotive control algorithms and electrical systems in the CRRC Zhuzhou Institute from 2010 to 2013. In 2014, he stayed for one year at the Hong Kong University of Science and Technology, as a research assistant supervised by Prof. Ming Liu, doing research on VLC-based indoor localization and 2D-laser based 3D sensing.

Interests
  • Robotics and Autonomous Systems
  • Artificial Intelligence
  • Robot Learning
Education
  • Ph.D. in Robotics and Artificial Intelligence, 2018

    Australian Centre for Robotic Vision (ACRV), Queensland University of Technology (QUT)

  • B.Eng. in Automation, 2010

    East China Jiaotong University (ECJTU)

Selected Publications

Adversarial Discriminative Sim-to-real Transfer of Visuo-motor Policies. In IJRR, 2019.
Modular Deep Q Networks for Sim-to-real Transfer of Visuo-motor Policies. In ACRA (Best Paper Finalist), 2017.
Tuning Modular Networks with Weighted Losses for Hand-Eye Coordination. In CVPRW, 2017.
The ACRV Picking Benchmark: A Robotic Shelf Picking Benchmark to Foster Reproducible Research. In ICRA, 2017.
Let the Light Guide Us: VLC-based Localization. In RAM, 2016.
Towards Vision-Based Deep Reinforcement Learning for Robotic Motion Control. In ACRA, 2015.

Projects

Learning Real-world Visuo-motor Policies from Simulation

Learning Real-world Visuo-motor Policies from Simulation

This is my Ph.D. project in the Australian Centre for Robotic Vision at QUT, with supervisions from Prof. Peter Corke, Dr. Jürgen Leitner, Prof. Michael Milford and Dr. Ben Upcroft.

Robotic Manipulation for Warehouse and Household Applications

Robotic Manipulation for Warehouse and Household Applications

Amazon Picking Challenge (2016) As part of the Team ACRV for the Amazon Picking Challenge 2016, I worked on hand-eye calibration with Dr. Leo Wu. Household Applications This is a project I worked on during my visit to the Perception and Robotics Group at the University of Maryland, College Park, Sep-Dec 2016.

VLC-Based Indoor Localization

VLC-Based Indoor Localization

This is a project I worked on with Prof. Ming Liu and Mr. Kejie Qiu when I was a research assistant in the RAM-LAB at HKUST. Contributions: Developed a beacon code selection algorithm and a decomposition algorithm for blindly mixed beacon signals, based on CDMA code selection principles and Gold-sequence correlation properties.

Academic Service

  • Membership: IEEE Robotics and Automation Society
  • Journal Reviewer:
    • RA-L
    • T-ASE
    • TNNLS
    • ISJ
  • Conference Reviewer:
    • ICRA: 2017, 2018, 2019
    • IROS: 2017, 2018, 2019
    • ICIA: 2016, 2018

Contact