A theory for optimal deterministic characterization of time-varying human operator dynamics

A deterministic theory of characterization is presented which can be used to determine the time-varying dynamics of the human operator engaged in a tracking task. With this theory it is possible to obtain a time-varying impulse response function and a time-varying transfer function which represent the action of a human operator in an open- or closed-loop control system. No special form of input is required. The characterization, that may be in either real-time or nonreal-time, is based upon an exact theory of fixed-form optimization. A strongly convergent, definitely stable iteration technique can be used to realize the optimal characterization filter. The theory takes the time variation of the impulse response or transfer function into account, so that it is unnecessary to make the assumption of slowly varying dynamics. An uncertainty or compromise exists between the error, i.e., the error between the output of the human operator and that of the optimal characterizing filter, and the degree of variability of the optimal characterizing filter. This uncertainty is fundamental, and therefore cannot be circumvented. Although the theory has been verified by extensive experimental study, emphasis here is placed upon presentation of the theory.