IJCAS

System identification is a field of control engineering that deals with the preparation of a mathematical description by recognizing the static and dynamic properties of automation systems. It becomes particularly important in the black-box approach, in which the modeling technique constructs a model using only the output data obtained from the system based on the known input signal. One of the most complete and powerful identification methodologies available today for the identification of nonlinear systems is the NARMAX approach. This paper presents and compares three methodologies used to approximate the unknown structure of a dielectric electroactive polymer actuator by applying one-step and multi-step prediction. The motivation of this study was to check the possibilities of the recent identification techniques on the object with complicated dynamics like DEAP actuators.

Keywords: DEAP, dielectric electroactive polymer actuator, machine learning, system identification.