This is 1346’s first time using mecanum wheels and we’re having some difficulty with the controls.
We have the vector addition equations and everything but currently working on a closed loop system using optical encoders. Is this the best way of going about it?
For those of you who have built mecanum drive trains before: What is the best way to “close the loop” and get feed back?
Gyro? Encoders? Other techniques?
Any comments on the programming is also highly appreciated cuz I am the programmer after all…
Actually assuming you have the equations working properly, that’s the trickiest part. Encoders will give you the speed of each wheel which is what you need to control fwd/back/strafe etc. For twist you need a gyro. Now implement a simple proportional control loop for each wheel and it should drive like a charm.
Most holonomic drives don’t behave AT ALL without some sort of feedback. Although sometimes if you just use a gyro (without encoders) and some clever programming, you can get by.
We are working on the mecanums. I’m using a combo of the optical encoders made for banebots gear boxes and the gyro. Not too sure how it’s going to turn out. I’m writing code now, I’ll test it tomorrow. I’d be happy to share when I figure it out.
The drive train isn’t usable without any feedback. When commanded to drive straight, it moves in a kind of arc and I bet if the floor is big enough it’d do a perfect circle (probably motor bias).
Recently we broke an encoder so it’ll be a few days until the new one ships in but in the meantime I can always to figure out how to implement the gyro.