Performance Prediction in a Single- Operator Simulated Surveillance

A semiautomatic radar surveillance system was simulated using a time-compressed, real-time cathode-ray tube display. Subjects were required to detect targets entering the surveillance area, initiate automatic tracking of these targets, and reinitiate lost tracks when automatic tracking failed. A within-subject response surface methodology (RSM) central-composite design was employed which permitted simultaneous investigation of the effects of five system parameters on surveillance operator performance. Response surface fits (second-order polynomials) were obtained and analyses of variance were conducted to describe these effects on two dependent measures of performance. Results support the contention that operator performance may be dependent upon complex relationships among the five system parameters tested. Furthermore, an RSM central-composite design provided an efficient method for obtaining data and quantifying these relationships