Self Stabilizing Chimney Climber

The first RCX climber was slow and unbalanced. The second NXT version was much faster, but still not very stable since it used the same techniques. This third version is a new design, made so that it will automatically balance itself on one axis. Balance on the other axis is achieved with an acceleration sensor.
Imagine a line from the left wheels to the right wheels, through the middle of the robot. On this line, the robot tilts when it gets unbalanced. Unlike in the older versions, the main weight is now below this line. This means that it will keep itself balanced without help.
However, there is another possibility for the climber to crash. That happens when either of the sides loses a bit of grip or if both sides do not go as fast as each other. You can correct this error by making one of the sides go slower, so it will achieve balance again. You would only have to know how big the error is, and in what direction. This is what the Mindsensors Acceleration sensor measures. A relatively simple NXT-G program keeps track of the sensor and motor speeds. It goes up until it sees the ceiling and then comes down again.


Matstver said…
wow i realy like it and sins its nxt i can build it to=)
Anonymous said…
This website is amazing, you really should add adsense to it because you could make so much money
Anonymous said…
1. with a bit of technology and good marketing, this thing might score the new household hit! a fullyautomatic chimnycleaner!
2. i suppose developping a robot like this takes a lot of testing and development, did you used safetymeasurements for the prototyping?
Laurens Valk said…
Thanks! Sorry for the late reply, I missed your post. There are two ways you can look at it. In total, it didn't take me very much time. However, this thing had two predecessors, which you could count as developping time :).

That doesn't mean I got this design at the first try of course. I first tried a different wheel system, but it turned out they did not have enough grip. Then, I had a different design of the base. The NXT was mounted in a very different way, so that it would lower when the thing was going up, without a third motor. This didn't work either due to a weight and grip problem, which I could explain to you if you're interested.
Then I came to the final design, the one you saw.
Actually, I didn't use any safety measurements :). Ever since this thing came off the ground, it was very stable. In cases it would fall down, you could clearly see that even 20 seconds before it eventually fell, so I had plenty of time to jump in.

Popular Posts