About
Li-Wei’s research interest lies in robotics and reinforcement learning. To build robotics systems, Li-Wei has been trained in various subjects including Machine Learning, Automatic Control, and Computer Vision, etc. Funded by the Ministry of Science and Technology, he works on an undergraduate project to create a patient-adapting oral model for robot swabbing in high-fidelity simulation. His enthusiasm earned him the Presidential Award while taking several upper-division courses and master-level courses. His interest in robotics stemmed from the movie, Prometheus. The scene where Dr. Shaw use the medical pod to do a cesarean section impressed him with the robot’s dexterity and intelligence. To stay abreast of state-of-the-art technology, Li-Wei has participated in Taoyuan ROS summer school 2021 and Nvidia GTC 2021. He is currently exploring reinforcement learning in Isaac Gym and ROS.
Center for Artificial Intelligence and Advanced Robotics
Feb. 2022 — Oct. 2022
Research Assistant
§ Topic CHARM-Companion Healthcare Aid Robot Manager
§ Skills Android, Java, Kotlin, SQLite
• Connected an ECG smartwatch to the robot using BLE protocol and utilized SQLite to synchronize multiple users' physiological data to a remote server.
• Utilized synchronous and asynchronous threading to develop a surveillance module that follows and recognizes faces and interacts with people in voice to improve user experience.
• Refined research proposal and informed consent form, which helped the research pass Institutional Review Board and advances into the clinical research phase.
Robots and Medical Mechatronics Lab
Jul. 2020 — Mar. 2022
Undergraduate Researcher
§ Topic Swab Robot, Field Robot.
§ Skills ROS, Python, SolidWorks, MATLAB.
• Developed a statistical morphing oral model that fits the oral cavity in 3D Slicer and built a simulation environment in GAZEBO for the operator to swab for the oral specimen.
• Won sponsorship from Taiwan's Ministry of Science and Technology (MOST).
• Designed the torque of counterbalance in MATLAB, broadening the robot's workspace by 80%