This little buggy was one of my first practical electronics projects. It was controlled by a high-end
basic stamp microcontroller (ick!), but interfaced with a nifty GPS unit by Garmin which had a really
convenient serial data output port on the back.
Basically, you could program into the GPS an arbitrary route consisting of waypoints, and the
microcntroller figure out how to get the car there using a rudimentary proportional control
algorithm that can best be described as "highly underdamped". However, it worked. Well enough, at
least, that you could set this thing down in a large field, let it go, and it would drive out of sight
while you sat the sat there waiting for about 5 minutes. Then it would suddenly reappear out of the
distance which it's high pitched motorized whine, having traversed the path assigned to it. Of course,
it didn't do any sort of obstacle avoidance, so it wouldn't work it an area with lots of trees or other
such hazards. But in large open areas, it was hilarious.
Here, you can see the Garmin 12 GPS unit on top, and the basic stamp mounted on a dev board on the bottom. The power sources on this vehicle were somewhat ridiculous -- it used four seperate supplies. One beefy 7.2v pack powered the car, four AAs powered the servo controller, another four AAs ran the GPS unit, and a 9v supplied the basic stamp regulator.
Version 1.0. Yes, it's embarrasing to admit it -- that's the wireless radio control transmitter taped to the car. Two servos strapped to the controller actuated the throttle stick and the steering wheel knob. Pretty pathetic! But it also, actually, worked. I did this because at the time I didn't have speed control set up, and I wanted to test the control algorithm & GPS interface. It failed once when the servo which worked the steering wheel knob bounced out of the car, leaving the vehicle on a doomed trajectory, ultimating driving full speed into a ditch. But that was worked out in version 2.0...
With the success of the ground-based vehicle, I had hoped to continue the project and modify it for use as a general purpose autopilot system for model airplanes. However, time got the better of me, and difficulties with the pressure sensor I was using as barometric altimeter forced me to put the project on hold, indefinitely, rather than risk destroying a great model aircraft.