PID controller implementation written in C.
Note on 'derivative-on-measurement': Since the 'error signal' effectively going into the differentiator does not depend on the setpoint: e[n] = 0 - measurement, and therefore (e[n] - e[n - 1]) = (0 - measurement) - (0 - prevMeasurement) = -Kd * (measurement - prevMeasurement). (Note the minus sign compared to derivative-on-error!) I've included the minus sign in the code, so gains will have the effect as normal.