Extracting fuzzy control rules from experimental human operator data

This paper proposes an approach where the interpretation of manual control strategies is carried out by modeling the human operator as a fuzzy logic controller. The linguistic rules thus obtained can provide a better insight into the operator's actions, allowing mistakes to be more easily pinpointed and corrected. Instead of extracting the control rules directly from raw experimental data, an intermediary ARMA model for the operator is employed to improve the data consistency. For illustration, this method is applied to the problem of supervising an apprentice operator, with basis on rules extracted from the actions of an experienced manual operator.