Education

• M.S./Ph.D. in Electrical and Computer Engineering, University of California San Diego (UCSD), 2019 to 2024 (expected)
• B.S. in Electrical Engineering, National Taiwan University (NTU), 2013 to 2017

Work experience

• Research Assistant from July 2017 to June 2019 @ Center of Artificial Intelligence and Advanced Robotics (AiROBO), National Taiwan University
  1. Improved a self-conscious-based decision model for Pepper
  2. Developed an intelligent secretary system for RoBoHoN (GitHub)
• Part-Time Student Intern from July 2016 to December 2016 @ Mechanical and Mechatronics Systems Research Laboratories, ITRI, Taiwan
  1. Developed virtual reality functions on Oculus
  2. Implemented a simultaneous localization and mapping (SLAM) algorithm in Gazebo

Publications

Please check out the publication list here.

Research Projects

• Motion Planning for Autonomous Suturing
  • Advised by Professor Michael Yip @ ARCLab, UCSD
  • Developing and testing an autonomous suturing framework, including perception, planning, and control
• Suture Needle 6D Pose Tracking (Paper, Video)
  • Advised by Professor Michael Yip @ ARCLab, UCSD
  • Tracked the 6D pose of suture needles with particle filter under different scenarios, including ex vivo environments.
  • Proposed three observation models based on the circular geometry of suture needles, and derived the noise variance for one of the models.
  • Experimental results demonstrated that our proposed method outperforms previous ones and is robust under occlusion and complex backgrounds.
• Bimanual Regrasping for Suture Needles (Paper, Video)
  • Advised by Professor Michael Yip @ ARCLab, UCSD
  • Proposed ego-centric state/action spaces and mixed exploration strategy for reinforcement learning (RL) to learn a generalizable policy for suture needle regrasping
  • Achieved 97% success rate in simulation and 73.3%/90.5% success rate in real-robot experiments
• Parallelized Reverse Curriculum Generation (Paper)
  • Advised by Professor Hung-Yi Lee and Professor Li-Chen Fu @ AiROBO, NTU
  • Improved reverse curriculum generation by training multiple actor-critic (AC) pairs and swapping their critics periodically
  • Achieved better convergence and generality, and the proposed method is applied to several robotic manipulation tasks
• Robots Learn Hand-Ear Coordination (Report)
  • Advised by Professor Li-Chen Fu @ EE Department, NTU
  • Implemented Deep Q-network (DQN) and prioritized experience replay for a robot to react to musical tempo from noisy music
  • Achieved 97% accuracy of end-point detection by dropout and L2 regularization
• Avatar: Robot Imitates Human Motions in Real Time using Kinect 2 (Report 1, Report 2)
  • Advised by Professor Li-Chen Fu @ EE Department, NTU
  • Mapped the skeleton data, extracted from Kinect 2, and robot postures by (1) direct joint mapping and (2) supervised learning

Term Projects

• Automatic Piano Accompaniment Robot (Report, Video)
  • Advised by Professor Li-Chen Fu @ CSIE Department, NTU
  • Used LEGO EV3 to build a robot, and implemented retrieval by singing and humming as well as trajectory planning in Matlab

Honors and Awards

• 2022 Grace Hopper Celebration Student Scholarship
• 2022 Inclusion@RSS Program Fellowship
• 2016 The 24th Conference on Automation Technology College Student Paper Contest, Second Place
• 2016 Innovate Asia Design Contest Outstanding Achievement Award

Academic Service

• Program Committee, ICLR 2022 Workshop on Generalizable Policy Learning in the Physical World
• Reviewer, IEEE International Conference on Robotics and Automation (ICRA)
• Reviewer, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Skills

• Programming Languages: Python, C++, Matlab, Android Development, Java, System Verilog
• API and Toolkits: Tensorflow, PyTorch, CoppeliaSim, Gazebo, PyBullet, OpenCV, OpenPose, Oculus, LaTex
• Operating System: Linux, ROS, NAOqi, Windows
• Languages: Mandarin (native), English (proficient), Japanese (basic)