I've made two incursions into the world of Lego Mindstorms robotics and both times very narrowly averted falling into total, permanent addiction.

Click on an image to see a larger version, and use your browser's "Back" button to return here.

Formula 1

This low-CG wheeled 'bot was my first fully autonomous (self-contained, self-controlled) creation. It was designed to drive itself around a racecourse using side-mounted light sensors to detect the edge of the track. Though I never had a large enough area to build a complete track for it, it really worked!

Click here to see the test run!
(.MOV file - requires Quicktime)

In the video, it's running at 30% speed for the sake of safety.

The Racers

I really liked the F1 car and decided to double up my efforts for the next round of creations. I came up with a "competition" with myself to see which of two robot designs would be the most efficient at driving around a track. The first design (with #2 on top) used independently driven wheels for propulsion and steering and a "dumb" caster at the rear. The second design was rear-wheel-drive with one motor and a differential, and had a live front wheel strut for active steering.

The second design worked better, but this was mostly due to the forward mounting of the sensors. Overall, though, I liked the first much better. It's very simple, reliable, and agile. With more sophisticated programming, I could have made it into a formidible navigator as well.

The Walker

This was the first in a short series of walking 'bots I experimented with. This one would have needed more gear reduction to have enough strength to carry its control module & batteries, but it was fast and agile (forward/reverse/turn) and fun to control remotely. My first working autonomous walker used eight legs and had a simple obstacle-avoiding program. It could also be controlled by infrared remote.

Sparky

This is where things started really getting out of hand. Sparky was a mostly autonomous pet designed to play fetch with a yellow ball. To see, he used the Vision Command color camera, which was attached to my PC via a USB cable. He could look for the ball and go after it, using his curved side "wiskers" to avoid running into things along the way. The wiskers also helped him feed the ball down to his grabber hand once he found it and drove up to it. Propulsion & steering was accomplished via independently driven wheels at the sides.

When it was time to pick up the ball, the grabber opened a motor in the aft section tilted the rear up, making Sparky lean forward. Check out the grabber action:

He'd then drive forward a bit and grab the ball, and lean back. A touch sensor in the grabber hand would confirm if he had actually gotten the ball, since he couldn't see straight down. Then he was supposed to turn around and bring the ball back, but I had trouble with that part (even robotic dogs can be ornery!).

The most fun thing was that he was programmed to sit in one spot, but get excited and go into fetch mode when you'd show him the ball. When he was excited, he'd wag his purple tail and gitter around a bit.

Sparky used two "RCX" computerized control units to handle his many functions, and the two would communicate between eachother, like a left & right brain, via IR.

By the time Sparky was built, I was already into R/C, with the Phase I Ultimate Rustler project underway, and I decided to choose just one hobby and sell off the vestiges of the other.

R.I.P. Sparky.