Chaos-Entropy Analysis and Acquisition of Individuality and Proficiency of a Human Operator's Skill Using a Fuzzy Controller

Abstract

In the present paper, we use time series data and fuzzy inference to identify individual differences in the behavior and proficiency of human operators performing various skills and use the identified individual skills as a fuzzy controller. Human operators are trained to a certain skill level at stabilizing an inverted pendulum, and the data obtained in 10 trials per operator were successively used for analysis, where the waveforms of pendulum angle and cart displacement were analyzed. The maximum Lyapunov exponents were estimated from time series data with respect to embedding dimensions. The fuzzy controller identified from the time series data for each trial and for each operator represented well the human-generated decision-making characteristics, exhibiting chaos and a large amount of disorder The estimated degree of freedom of motion increases and the estimated amount of disorder decreases with the increase in proficiency, both in the experiment and in the fuzzy control simulation. It is also revealed that the agreement between the experiment and the fuzzy control simulation for the degree of freedom of motion and indicates that the entropy ratio is particularly good when the measured waveform and the simulated waveform are similar in appearance.