This is a quick demonstration on one of our older robots of our autonomous bridge balancing. We know it doesn’t have to be autonomous, however, it can balance faster than a person most of the time.
Here’s the link:
what kind of logic are you using?
Do you think it is more useful balancing autonomously?
To me it doesn’t seem like balancing autonomously would not be as stable as a human especially with more than one robot on the bridge.
In theory, the balancing algorithm will work even if there is another robot on the bridge!
Nicely done!
A simple pre-programmed backwards motion (with each robot accounting for it’'s appropriate shift on CoG), when properly calibrated, could be run by many roots on an alliance and everything would work perfect. If only one robot was driving, you would have to compensate for the extra weight of the other robots, but it shouldn’t be all that bad. Maybe something other than a pre-set bakwards motion could account for different distributions of weight.
Normally I would say humans should be able to do this better, but that program is beautiful :).
To be useful, the routine would have to have a “pause” that leaves one side still touched down while a partner joins it. If you pause too soon, there might not be enough room for the partner. Too late and the bridge tips up going to balance without the partner.
imaginary two-robot sequence.
alliance agrees which two robots will attempt balance on their bridge.
also agree which is balancer and which is “passenger” B and P henceforth.
both B & P to approach from the same side.
B lowers and mounts bridge advancing toward center.
B pauses with enough room for P.
P follows onto bridge until fully supported by sloping surface.
P stops and “brakes” if needed.
B resumes auto balance mode by driving past fulcrum to get both balanced.
(problem if B cannot move far enough away from P to balance bridge. is this possible? likely?)
We are using two gyros that give us the angle of the bridge. If the angle changes sign we go the other direction and a bit slower. This is 100% student written.
To all other posters. It is important to note that we don’t know the CoG when multiple robots are on the bridge, but the algorithm has been tested to work even then.
We are planning on being last on the bridge and having our team mates as far forward as possible.
I will be happy to answer other questions if you have any!
Anybody auto-balancing by using a changing center of gravity within the robot, e.g. battery on a sliding rail?
That is too complicated. Keep it simple. After both robots are driven on the bridge then engage the balancing algorithm. It should work the same as if the robot were there alone (provided the weights are not too different).
Looks good!
That is in the works on the mechanical side of things already although I don’t want to reveal too much.
That is exactly right. We have this routine set to run with the push of a button. We don’t need the over-complicate the algorithm if we can start it whenever we want. Also because we are planning on being full weight, we shouldn’t have too much of a problem balancing with others.
Something like that. Combined with gyros.
Easy to keep it simple if all the processing is being done by one brain. A little more complex even if one team (two drivers?) has practiced the maneuvers with a single robot for a while. still more complex when doing the balancing with a (random?) alliance partner for the first time. Yep! complexity increases again if the alliance partner is on the other side of the ball corral. Then as if icing on a cake, increase once more if the balancing robots’ drivers are on opposite sides of the field.
Good to have the choreography down before the dance begins, eh? 
P.S. I probably won’t get to elims, so why think about three on a beam? :yikes: