Low order models for human controller - Mouse interface

Abstract

Modeling human operator's behavior as a controller in a closed-loop control system finds applications in different areas such as training of operators by expert operator's model, tele-operation or developing warning systems for drivers. In this paper, first, an experimental setup has been developed for collecting data from human operators as they controlled a process with a DC motor, using the mouse interface on the Matlab environment. Low-order ARX models are proposed for human operator modeling. Replacing the operator by a stand-alone human controller model was one of the validation methods. Experimental results are shown to evaluate the performance of the proposed approach for human low-order modeling.