William Y
- Research Program Mentor
PhD candidate at University of Pennsylvania (UPenn)
Expertise
Computer Science, Machine Learning, Robotics, Control Algorithms
Bio
I am a third year PhD candidate at the University of Pennsylvania GRASP Lab, where I study control algorithms for quadruped and bipedal locomotion robots advised by Professor Michael Posa. I graduated with my BSE degrees in Mechanical Engineering and Computer Science from the University of Michigan in 2018. I served as the lead TA for two graduate courses, Control and Optimization for Robotics and Mechatronics, teaching over 200 students and writing a new autograder. I have also mentored several students in my own research, including 3 sophomore undergraduate students and a MS student in robotics. My expertise areas include machine learning, optimal control algorithms, hardware, and robotics. In my free time, I enjoy biking, backpacking, and climbing. I love being outdoors and will play almost any sport, except for basketball - I'm terrible at basketball.Project ideas
Optimal Control Methods for Simple Robot Systems
In this project, you will learn about and implement an optimal robot control algorithm for simple robotic systems, such as a cart-pole or 2D quadcopter. We will begin with a brief survey of state-of-the-art control methods, develop a strong understanding of the underlying physics, and then implement the control algorithms in a Python simulator to validate our methods. This will include a guided exploration of advanced controls techniques - Python experience preferred.
Review of Legged Robot Morphologies
Given the extensive variations in quadruped robot morphologies in both research and commercial platforms (think Boston Dynamics Spot vs ETH Zurich ANYmal), let's explore how the different hardware designs affect performance and types of tasks. This project will involve a review of the hardware design decisions from a selection of some of the most influential quadruped platforms and an analysis on how different configurations affect the robot dynamics.