University of Michigan
Robotics
University of Michigan
Robotics
University of Michigan
Robotics
HELLO, I'M
KEN TAI
Tzu-Chieh Tai
University of Michigan
M.S. in Robotics
National Tsing Hua University
B.S. in Mechanical Engineering & Electrical Engineering
Interdisciplinary Program of Engineering
Gyro System Inc.
Intern
MY BACKGROUND
About
I am passionate about the intersection of robotics control and AI perception. To me, the future of robotics lies in end-to-end reinforcement learning guided by vision—the same way we humans sense and react to our environment. I'm on a mission to make robots more intuitive and adaptive. Explore my journey on this page!
Selected Projects
Mobile Robots
Autonomous Mobile Robot Exploration with LiDAR using MBot
• Engineered a robust scan buffer to manage asynchronous LiDAR data streams, ensuring continuous data
availability for SLAM and scan matching, achieving a mapping error under 5cm.
• Developed a custom scissor lift with a 2.8x extension ratio (5cm to 14cm), engineered specifically for
steady and balanced lifting of blocks during vertical manipulation.
Application is no longer available.
Rescue Robot with Autonomous Mapping and Object Detection Function
• Developed algorithms and combined sensor data to guide autonomous car to explore unknown
environments, achieving over 90% exploitation rate in real-world scenarios
• Integrated 2D and 3D data to provide comprehensive information on disaster areas, distributed
computational loading through ROS architecture to perform casualty detection and exploration
• Earned 3rd place and honorable mention in the poster exhibition and project competition contest
-
rescue robot performing autonomous mapping in unknown environment
-
Demo video
Robotic Arm
Robotic Arm Trajectory Planning
• Led the integration of the collaborative robot arm with Moveit! through ROS, built up digital
twinning model of the entire system using ROS and C++ for scenario testing and execution
refinement
• Discovered the obstacle avoidance issue in simulation and solved it by designing moving strategies
and paths to complete industrial pick-and-place missions with high precision
• Implemented reinforcement learning to optimize existing trajectories through building digital
twinning using Pybulle
Application is no longer available.
Vision-Based Manipulation with RGB-D Sensing and a 6-DOF Robotic Arm
• Lead the integration of kinematics, calibration, and coordinate transformations.
• Overcame a complete computation stall in depth image projection, achieving precise real-time
contour detection.
• Engineered precise block detection and achieved five-block stacking with a consistent error below
10 mm.
Application is no longer available.
Research
Trajectory Optimization of a Glass of Water
• Applied geometric analysis to optimize polynomial trajectories under acceleration limitation
• Developed methods to perform planar motion planning and applied simulations on fluid dynamics under motion, discovering cubic polynomials that effectively improve smoothness
• Represented the College of Engineering in the internal preliminary selection for the Chun-Tsung
Annual Conference
