Eliminating Deadband On Motors

[WizenedEE]

Has anyone thought about eliminating the deadband on the drive motors with an omnidirectional drive system? What I mean is, when the motor is given a speed of "0.1", it would give it something like 20% power so that the motor would actually go at 1/10 is maximum speed? I had to do this for a smaller robot I was making since I was using some pretty bad wheels, but I noticed while driving our team's robot today that it liked to go directly diagonally more than a little bit to either side.

Have any other teams seen the need to do this? Has anybody done this and saw an appreciable difference in driving? Are there any warnings for someone trying to do this, other than making sure to create a deadband on the joystick?

On a slightly unrelated note, does anybody know how to create a waveform chart in Labview that has a discontinunity, so that it will graph -255 to -75, 0, and 75 to 255 as if they were right next to each other?

[JewishDan18]

i'd use encoders on the wheels combined with a PID loop to do "speed control" on the wheels, so they rotate at the correct speed. Searching on "pid" will give you much better explanations than i can give

[lemiant]

As far as the graph, a simple solution would be subtract 75 from the absolute valu and then give it back it's sign (processed_Val*orig_val/abs(orig_Val)), before outputting it.