MATHEMATICAL MODELS OF HUMAN PERFORMANCE IN MAN-MACHINE SYSTEMS.
Abstract
The report describes three research approaches to the problem of mathematically representing human performance parameters in weapon, maintenance, and command and control systems. In the first approach, twenty operations research analyses and models of military systems were examined to determine if the models included human factors parameters and to what extent they were sensitive to variations in these parameters. Although many of the functions of the systems modeled were performed by humans, human performance parameters were not, in general, sufficiently defined to permit mathematical or empirical manipulation within a man-machine simulation framework. In the second approach, an attempt was made to establish predictive relationships, based on regression and factor analysis techniques, between human engineering design parameters and those criteria of systems effectiveness, such as maintenance task time, that can be transformed into a more molar index--system downtime. The human engineering predictor-parameters accounted for 50% of the criterion variance. In the third approach, a series of experiments involving real-time simulation of a command and control system was conducted to determine if, and how, a computer might aid diagnostic performance (in tactical decision making) in threat evaluations. The system output or criterion of effectiveness was the degree to which the system assesses the true state of threat. With computer aiding, correct decisions increased by 13%. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1968
- Accession Number
- AD0673348
Entities
People
- Donald A. Topmiller
Organizations
- Air Force Research Laboratory