PID controller tuning based on disturbance attenuation FRIT using one-shot experimental data due to a load change disturbance

Abstract

A direct design approaches based on input-output measurements with no need for help from a plant model have attracted attention from several researchers. We have recently proposed such a disturbance attenuation FRIT method using input-output data generated by disturbances. The approach has advantages that it can tune PID gains to improve feedback properties, such as disturbance attenuation. Furthermore, the method has been applied to PID control gain tuning. The method tunes PID gains using one-shot experimental data generated by a known step-type disturbance added at the input signal. However, additive known step-type signal implies a test signal to identify the characteristic of the control systems, so the case where the method can be applied may be restricted. The paper, therefore, gives a tuning method using a one-shot experimental data generated by a load change. Such a disturbance has a possibility that the data can be collected during a full operation. The proposed method can be realized by estimating the magnitude of an impulse and step signal for a load change disturbance from one-shot experimental data. The proposed method realizes the approach by modeling a load change disturbance as a linear combination of a step signal and an impulse signal, where the weighting parameters are unknown. The paper also gives the way how unknown parameters are estimated from the one-shot initial input-output data. Finally, this paper shows the efficiency of the disturbance attenuation FRIT through the experimental result of a helicopter attitude control model when the disturbance is known or correctly estimated.