My projects and works related to MIT or other real-world pursuits are here. Individual pages, if they exist, are linked through the images.
2.737 Mechatronics and ongoing personal project Segfault is a balancing scooter (Segway-esque) which runs a 100% hardware controller. That is, no code runs to keep the vehicle stable. All signals are processed with passive and op-amp based filters. The vehicle uses a fairly standard accelerometer and gyroscope complement, but with the processing occurring in continuous-time. I also built a set of custom Class-D amplifiers to power the drive motors.
2.009 Product Engineering Processes team final project. The year’s theme was food, and what better way is there to serve food than using a cluster of adorably cute disc-shaped robots? The Noribo (nee Sushibot) project aimed to create exactly this. The physical implementation was an exercise in rapid prototyping and PCB design, but another important part of the whole project, the business case, also had to be researched, developed, and finally integrated with the product in a final presentation in front of everyone. And by everyone I mean about 1,000 people from both MIT and industry. This was legit.
2.007 Design and Manufacturing I 2009 competition robot sculpture. Having more build experience than most in MIT’s sophomore design class, I tried to execute the one-week robot blitz at the end of term. It worked – except for one minor detail. The tragic tale of 2.007 is linked through the image.
6.115 Microcomputer Project Laboratory final design project, which caused me emotional trauma and my swearing off of programmable microcontrollers for the next 2 weeks. An attempt to make a hard drive clock turned into a flickering, sporadic hard drive something-or-other that crashed if the light bars tried to overlap each other or if one got near the zero mark. The class was programming-heavy, something I was not used to. Even worse? It was in ASSEMBLY LANGUAGE.
6.131 Power Electronics Laboratory final design project, recap part 1, part 2, part 3. For once, an academic project I am content with, since it actually did function as expected. The idea was to make a logic-gate commutated brushless motor driver, i.e. one that used no programmable microcontrollers (due in part to my lingering post-traumatic school disorder from 6.115). While a fun thought experiment and attractive at the time, I now know they make these things on a single IC.
2.12 Introduction to Robotics final project robot, which harvested PVC pipes and delicately ran MATLAB, Simulink, Python, and embedded Linux, all over WiFi. It worked until MATLAB couldn’t locate its license file.
Research & Teaching
2.007 Teaching Assistant & Lab Assistant I’m currently a recurring TA for MIT’s 2.007 Design & Manufacturing I class … yes, that class listed third up there which I wrote a pretty detailed summary on.
Electric Vehicle Design; or “2.00Gokart” In what is perhaps my longest running “project”, I’ve been the primary instructor for a design lab class focusing on the design and construction of small electric vehicles. Unlike almost all of these project type classes, no “kit” of materials is provided, and the students must compile bills of materials to be ordered. The goal of the class is to include supply chain considerations and resource management into the undergraduate design project curriculum, where it is not usually encountered. Additionally, students must learn to compromise or base designs around existing, commercially available parts. The resource base is the collection of vendors, parts, and techniques I’ve compiled over years of building these types of small vehicles. (Note: The current link goes to a summary post on the blog section; when I put together a more formal page for it, the link will change.)