PID controller improvement

[ushardul]

Can't use derivative gain currently because we are testing on step samples, so near an edge, there are very large changes, resulting in large corrections using derivative gain. Implementing a low pass digital filter on the error signal --may-- give better tracking for the step edges.

[neil-sarkar]

I think the step samples are precisely what we need derivative gain for... will try imaging and see how the PID settings affect our tracking performance

[ushardul]

Yeah I tried imaging with a very small derivative gain but it flattens the scan. Like, I said, I think a low pass digital filter on the error signal may give better tracking for steps.