Misc improvements (step_info, param optimization)

[bresch]

A couple of misc improvements I've been working on recently:

• allow negative feedforward gains (this could be useful to have more freedom and reduce overshoot in the step response)
• select between OLS (ordinary least-squares) and RLS (recursive least-squares)
• automatically try to optimize model order, delay and filtering params (button "find parameters). This can be slow
• expose pre-processing filtering parameters
• show fit quality based on normalized root-mean squared error (NRMSE)
• show if the identified model is stable. If unstable, suggest stabilization using pole mirroring.
• show step response rise time, settling time and overshoot. Also allow setting them as requirements and display on graph

[mahima-yoga]

Did a first pass. Didn't get to test everything yet - the parameter search is cool (as long as the data set is not too long).

[mahima-yoga]

Super cool