/CarND-PID-Controller

Primary LanguageC++MIT LicenseMIT

CarND-Controls-PID

Self-Driving Car Engineer Nanodegree Program

PID Selection Strategy

The PID coefficients were selected manually, by trial and error.

  1. I used only enough proportional gain in order to make the car responsive
    without oversteering in the presence of a small error.

  2. Then I used the derivative term to compensate for the curvature in the turns.
    I realized this resulted in much smoother driving than using the integral term.

  3. Lastly, I used the integral term to make the error compensation occur faster,
    I thought of it as sort of a "helper" of the P term.