Tuning proportional-integral-derivative controllers using achievable performance indices

This paper proposes a simple method to synthesize a proportional-integral-derivative (PID) type controller that is capable of providing the best possible performance for regulating stochastic disturbances and/or set-point tracking. This novel strategy requires no a priori information about the open-loop process or noise models. Input-output data from closed-loop experimental tests (e.g., through natural or induced set-point changes) will be used to estimate the closed-loop servo and disturbance models, which are then utilized to arrive at "optimal" PID controller settings. Several measures of the performance and robustness are computed alongside to aid in the process of controller tuning. The efficacy of the proposed method is demonstrated using representative case studies.