• Worked to combat the challenges of vehicle autonomy presented by weather

• Led development of confidential testing platform for internal use

• Utilized knowledge of mechanisms, sensors/actuators, and hydraulics/pneumatics in testing platform design

• Set up virtual lab station, designed and conducted experiments, and performed tests for data acquisition

• Implemented image processing system for detection using Python and OpenCV

• Liaised with senior engineers on Planning, Simulation, and Sys Eng teams to understand their work and

challenges

• Participated in discussions with the CEO, CTO, Engineering Directors, and Team Leads to learn about the

self-driving car problem and gain unique insights into current goals and approaches

• Hosted company-wide 30-min poster session and 1-hr final presentation with Q&A to showcase internship project

• Developed juggling control algorithm for a robot to repeatedly bounce a ball on its head while maintaining stability

• Programmed extensively in Matlab and familiarized self with lab’s existing codebase

• Implemented high level juggling controller to work with two distinct low level torque controllers

• Modified simulator and visualizer to include the ball-to-robot contact dynamics and ball flight dynamics

• Tuned algorithm on increasingly complex simulators before testing on hardware

• Utilized Agility Robotics’ Cassie, a bipedal robot with 20 degrees of freedom, as a testing platform

• Was the first lab to display successful juggling results on Cassie, achieving 42 consecutive bounces in experiment

• Published the work as an equally contributing first author at the European Control Conference 2020

• Video showcasing results (with > 25,000 views on Youtube):

https://hybrid-robotics.berkeley.edu/cassie-juggling/