Monday, March 28, 2011

Basic Design Philosophy

I've had a Logomatic for going on two years, and my main plan for it is to finally complete my college senior project, the "Rollercoasterometer". This device has three axes of accelerometers and three of gyroscopes. My senior project was building a fiber optic strain-gage accelerometer, but since then, MEMS technology has revolutionized the world and the sensors I described above now fit in two chips, a couple of milligrams and microamps each. Now, the device I imagined will fit into my pocket instead of a backpack. So, I bought a Logomatic and a set of analog MEMS sensors from Sparkfun, and put together the Mk1 Rollercoasterometer.

I've taken this device on roller coasters all over the country, from Disneyland to Disney World. It has gone through several revs, from one that fits into a small fishing box to one that fits into a much smaller box. However, I have never accomplished the full mission of the device, to actually measure the track. I have in mind carrying this device on something like Space Mountain, and running the recorded data through an integrator to get position and velocity information. I just haven't done it yet.

With the device being small enough to get into a small box (about 2.5x1.5x1 inch), new all-digital sensors, an additional 3-axis magnetometer as a compass, and a GPS, which made it the Mk4 "Rocketometer". I then packed it into the nose cone of a model rocket and launched it. Now when I integrate it, I would get the track of the rocket, from which I could get interesting things like peak altitude, top speed, and peak acceleration. This still didn't motivate me to write the filter, although I did write a pure integrator of the gyro data, to produce a movie of the orientation of the rocket.

I know that a Kalman filter, in particular a nonlinear unscented Kalman filter is the right way to do it. I just haven't been motivated to do it yet. Are you beginning to see what I mean by terminal ADD?

So, to get a deadline, to get into the local electronics community, to give me a project, to impress people, and TO ACTUALLY FINISH SOMETHING, I decided to enter the Sparkfun AVC 2011. My plan is to get a cheap but large RC helicopter, strip out its electronics and connect my Logomatic to the motors. Now I will have a deadline to actually write that Kalman filter, since it will be the navigation routine in a Guidance, Navigation, and Control (GNC) system for the robot.

No comments:

Post a Comment