A Discrete, Stochastic, Optimal Control Model of the Human Operator in a Closed-Loop Tracking Task.
Abstract
A discrete stocahstic optimal control model of the human operator is developed for the single-loop compensatory and pursuit tracking situations. The model generates signals corresponding to those in the physical closed-loop tracking situation. There is one primary model parameter which is varied to match model-experimental normalized tracking error at a bandwidth of omega sub B = 1.0 rad/sec for an input which approximates a rectangular spectra. With this parameter fixed, the model then predicts normalized tracking error and power spectra of control loop signals across a range of input bandwidths of 0.5 to 2.0 rad/sec. The model is applied to simple first- and second-order controlled elements in both compensatory and pursuit display situations. A comparison between model and experimental normalized tracking error and power spectral density data confirms the model capability of matching and predicting operator performance with sufficient correlation to warrant its application as a tool in manual vehicular control system design. Furthermore, the success of the model substantially confirms the hypothesis that the human operator behaves in some optimal manner when performing in a closed-loop tracking task. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1970
- Accession Number
- AD0719755
Entities
People
- Harvey M. Paskin
Organizations
- Flight Dynamics Laboratory