Teaching Robotics 

I've been a teaching assistant for 3 different robotics classes at MIT. They are all software classes that use the MIT Racecar platform for labs covering planning, perception and control. 

    6.141/16.405J - Robotics: Science and Systems 
       MIT's essential undergraduate robotics course. 

    MIT Lincoln Laboratory Beaverworks Summer Institute 
       A summer program for talented high schooler seniors. 

    6.a01 - Mens et Manus 
       A hands-on freshman seminar. 

As a TA, my role has been to write the labs, run the labs and hold office hours. For the most recent course, 6.a01, Andrew Fishberg and I were also in charge of developing the entire curriculum. 



6.141/16.405J - Robotics: Science and Systems  

This is the essential undergraduate robotics course at MIT. I took this course in 2017 and TAed it in 2018 and 2019. Students in 6.141 complete a number of labs that form the basis of a complete autonomous system: 

   - Wall following (PID, Ransac) 
   - Line following (color segmentation, homography) 
   - Monte Carlo Localization 
   - Path Planning (A*, RRT) 
   - Trajectory following (pure pursuit) 

The course culminates in an autonomous race and some sort of final challenge. The race looks like this: 

 
Students have a number of options for final challenges The year I took 6.141 I did autonomous parallel parking. As a TA I introduced "The Labyrinth" challenge. Student's were given the task of making their racecar escape a maze that is completely unknown in advance. Other options included high speed obstacle avoidance and deep visual navigation. My demo solution to the maze challenge looked like this. Some impressive student demonstrations are available here, here, and here. 


Beaverworks Summer Institute  

This is a summer program for high schoolers, hosted by MIT and Lincoln Laboratory, through their joint program Beaverworks. The course material is similar in structure and content to 6.141, although not nearly as in depth. Still, these kids pull of some pretty amazing results! The course ends with an autonomous race that requires various both lidar and vision based control systems (and switching between them!) 

Team 7 made a great video detailing the course: 

 



6.a01 - Mens et Manus  

6.a01 is a seminar intended for freshman who don't have experience with engineering. Half of the Racecar course is dedicated to modern machine learning techniques (neural networks) and how they can be used for vision based control.