H-Drive Question

[Amar Shah]

You could do two sideways wheels halfway along the sides of the U shape.

[z_beeblebrox]

Anyone have experience using a gyro to compensate for a center wheel offset from the center of mass?

[Nemo]

Originally Posted by z_beeblebrox

Anyone have experience using a gyro to compensate for a center wheel offset from the center of mass?

Our team did an offseason drive where the omni wheel was all the way at one end, so it had to compensate substantially in order to strafe straight. We used a gyro and managed to get it to strafe, but strafing was probably only 5 ft/s and we didn't stay with it long enough to get it to go really straight sideways. All we did was a simple proportional controller: we added an amount of turning to the joystick inputs / inverse kinematics equal to the gyro error times a tuning constant. Conceptually it isn't very complicated.

One of the things that does complicate life is that your strafe wheel isn't going to give you the same rate of angular acceleration as the drive does when it tries to spin the robot. That's what made our strafing a little shaky. I think that can be accounted for, but if those angular accelerations are different enough, it might take something fancy to really smooth things out.

If the center wheel was offset only somewhat from the center of mass, those problems get smaller, of course, and maybe you'd get by just fine with the simple proportional gyro feedback loop.

[Bruceb]

what will that center wheel to do robot balance when you are half way up the bump?
Seems it will get pretty wobbly.