For my job at Quickware engineering, one of the more nuanced problems I had to tackle was to create an Operator Control Panel that would provide an interface to the computer system we were replacing, but due to a byzantine pile of restrictions and client preferences, off-the-shelf software solutions were deemed untenable.
The solution we decided on was a fully custom lightweight CPU with a Harvard architecture that would execute proprietary assembly code stored in read-only FPGA memory. It would not be Turing complete and use instructions specific to our system, making the software impossible to compromise.
In order to prototype interactions, I built a fully-functional mockup of the system in Adobe Flash. When it came time to begin developing assembly code, I re-worked the mockup to act as an assembly code interpreter, so we were able to iterate the software design without requiring a re-synthesis of the board.
It's definitely a roundabout solution. But it is pretty cool.
I did some work on a multi-touch touchscreen with a couple of other Olin students. We were using the design popularized by Jeffrey Han at NYU.
Our screen could support multiple contact points, making hand gestures, multi-handed use, or multi-user interactions all very easy. Even better, the design was inexpensive and highly scalable. Not counting the projector or computer, total cost was <$300 for a 3'x3' display.
We were able to create similar hardware setup to Han's and get some prototype hardware going, but by the time we had cool stuff to show, the iPhone had come out and everyone just said "oh, so it's like a big iPhone." Bummer!
Still, it's cool to have been ahead of the curve for just a little bit.
In Fall 2006 in my Human Factors and Interaction Design course, I worked with a team of three other students to develop a new interface for Olin's poster printer. The HP drivers that ship with it are confusing, and a seven-page manual is placed next to the printer to help new users figure it out.
After evaluating how people use the printer, what their priorities are, and a careful assessment of design goals, we created a potential new interface that emphasizes intuitiveness and accurate real-time previewing. We then developed a mockup of the system in Macromedia Flash. For full details, head to our project site.
In Spring 2005, Douglas Bry of Wild Berry Productions commissioned by roommate and I to build a device that would allow a regular R/C airplane remote controller to control the pan, tilt, zoom, and focus of an HD digitial video camera that would be mounted on a jet airplane to take footage of other R/C craft.
Built around the PIC18F microcontroller and layed out on a custom PCB, our device accepts digital servo-style input from a standard speed controller. It controls pan and tilt through a hacked security camera turret-style base as well as zoom and focus on the camera over the standard camcorder LANC protocol.
I built an ultrasonic positioning system with a team of two other students sophomore year for Olin's Principles of Engineering course. The only requirements for the project were simply that it include a non-trivial electronic system and a non-trivial mechanical system.
Based around a PIC18F microcontroller, the ultrasonic positioning system offers tracking in three-dimensional space in an cube of approximately two meters in size, with accuracy down to about 1 centimeter, and can dump data to the computer over a USB interface with an update every 15 milliseconds, and could be built for a cost of less than US$100.
For more details, head over to the project site.