Book Description

The research investigates the feasibility of modeling human performance errors in application to the reliability analyses of man-machine systems. The research addresses itself to time-continuous tasks with the derivation of a general mathematical model of the probability of errorless performance which is equated to human performance reliability. The application of this model and the implications of the time to first error concept were tested with a laboratory experiment using a vigilance task. The observed times to first miss error, times to first false alarm error, and times to first combined miss and false alarm errors were ordered and, through classical interference theory, the underlying density functions were isolated. A number of distributions were tested for goodness of fit with the data. The Weibull, Gamma, and log-normal distributions emerged as relevant paradigms. The normal and exponential distributions were rejected. It was concluded that the derived general mathematical model of human performance reliability and the expected value of the random variable, time-to-first-human-error, are meaningful ways to quantify human performance of time-continuous tasks. (Author).