Heng Zhang (Andy)

张 恒

Control Engineering

Graduate with Masters

rail @ TJU

H416, Hua Building
College of Electronic and Information Engineering
Tongji University
4800 Cao’an Highway, Shanghai 201804, China

Email: hengzhang001@tongji.edu.cn



"Nothing in life is to be feared; it is only to be understood." --Marie Curie

About Me

I am now a graduate with masters from RAIL Lab at Tongji University (TJU), supervised by Prof. Qijun Chen. Before that, I worked as an assistant engineer in nuclear power plant design at SMNPC at China National Nuclear Corporation (CNNC). And I got my B.Eng. degree of Automation at Northeast Electric Power University (NEEPU). I have been lucky to work with Prof. Bo Li .

Highly-motivated in robotics with good foundations of math and programming. Proficient in data modeling and analysis, and enthusiastic about computer and the corresponding interdisciplinary subject. I'm interested in Robot Learning, Intelligent Motion Control, Cognitive Learning and their applications towards Artificial General Intelligence (AGI). My current research interest focuses include:

Research opportunities: I am happy to cooperate with others and/or answer questions about my research & associated academic programs. If you are interested, please send me an email. I am especially willing to cooperate with others who want to involve exciting projects targeting top-tier conferences/journals to reach out.


News & Updates


Selected Publications [Google Scholar]

A Survey on Imitation Learning for Robot Manipulation
[under review ]
Heng Zhang, Xianyou Zhong*, Zhengang Huang, Yuan Zhao, Chengju Liu, Qijun Chen (*=corresponding author)
IEEE Transactions on Cognitive and Developmental Systems@ (TCDS).
Sensor-Free Method with BP Network to Achieve Drag Teaching on the 7-DoF Collaborative Robot
[accepted ]
Heng Zhang, Xianyou Zhong, Zhengang Huang, Chengju Liu, Qijun Chen* (*=corresponding author)
China Automation Conference 2021 @ (CAC2021).

Intern & Work Experience

Project: Visual-based real-time robot arm fine-tuning imitation from human

  • github: my Repo.   Time:Feb. 2022
  • Sensor: Orbbec 3D camera , official SDK
  • Robot: a 7 DoF collaborative robot at our lab.
  • Work: The video presents my experiments: The position of my right wrist is percepted by camera and then the position error of the key-point during several response cycle of camera caused by human wave are sent to robot controller. The experiments ware conducted in our physics simulation environment, which was developed by Qt_5. Experiments on a physical 7 DoF robot is coming soon.

  • Project: RL-based implementation on robot manipulation tasks (Jan. 2022)

  • github: my Repo.
  • Experiments: Open AI Gym (MuJoCo)
  • robot: CRRC7, physicial simulation (in MuJoCo) of a 7 DoF collaborative robot at our lab.
  • Work: Robot can manipulate objects in different task settings, such as "reach","push","pick and place"... The experient is based on HER (DDPG). Plus, SAC algorithm can help the learning process of the agent. The video is coming soon.

  • China Railway Rolling stock Corp (CRRC), Tsingtao, China (Jul. 2020 - Oct. 2020)

  • Position: Research Intern in Control & Robotics Group
  • Mentor: senior engineer Yufu Hao
  • Project: High performance universal robot control platform
  • Work: dynamics modeling and drag teaching based on Sensor-Free method.



  • Project: Robot force control

  • Participator in Robot Arm Group @ TJU.
  • Mentor: PhD candidate Xianyou Zhong
  • Project: Force control research & implementation on the collaborative robot
  • Work: Drag teaching in high precision based on F/T Sensor using filter algorithm.

  • Honors and Awards


    Community Service

    Community service makes me have the spirit of dedication team-work, I am willing to make more contribution to the community.

    TA in Linear System Theory and Design
    One-star volunteer of China Foundation for Poverty Alleviation
    Student journalist in Alumni Association
    Class Monitor

    Software & Patents

    Control Method, Device and Equipment of Collaborative Robot Drag Teaching Based on Motor Current
    Qijun Chen, Heng Zhang, Chengju Liu
    CN 112894821A.
    [Patent]

    An indoor navigation method for Pepper robot
    Chengju Liu, Qijun Chen, Liwen Lu, Jiayuan Du, Heng Zhang,
    CN 113029143A.
    [Patent]

    Motion recognition of human skeleton based on lightweight graph convolution based on channel attention
    Chengju Liu, Ronghao Dang, Qijun Chen, Heng Zhang
    CN113111760A.
    [Patent]

    A text feature construction method based on Word2Vec and syntactic dependency tree
    Qijun Chen, Qiuchen Wang, Chengju Liu, Heng Zhang
    CN113111653A.
    [Patent]



    Acknowledgements

    I am very grateful to my advisor Prof. Chen and other teachers, my friends and classmates, who are Xianyou Zhong, Haoran Sun, Liang Tang, Zhengang Huang, Yuan Zhao, Jing Xu, Figueroa etc. Thanks to those who have helped me.


    Last update: Aug. 6, 2021