Abstract
In this paper, the human operator's learning processes are classified into the learning-memorizing process and the learning process, in accordance with whether the random appearing forcing function is the same on each trial or not, respectively. Under this classification, the transitions of the operator's model parameters in the two processes are analyzed and compared. The model is composed of gain, time-delay, neuromuscular lag and first order lead and lag. The model parameters are found to vary between the learning-memorizing process and the learning process, with the input bandwith and with the operator's practice. Some of these results are compared with those of the cross-over model. Furthermore, the accuracy of the proposed model is substantiated by checking the agreement between the model output and the operator's output by the correlation coefficient and the power ratio.
