For the development, optimization, and comparison of controller-models for the human operator a systematic method is proposed. This is suitable for any designed model, even multi-parameter systems. A random search technique is chosen for the parameter optimization. As valuation criteria for the quality of the model-development the criterion function, the comparison between the input and output functions of the human operator and those of the model, and the most important characteristic values and functions of the statistical signal theory (mean values, auto-and cross-correlation functions, histograms, and power density functions) are used.
Abstract