Crewcut: A New Tool for Predicting Human Performance in Conceptual Systems

Abstract

CREWCUT was developed in response to the need for a tool that can predict the effects of automation on operator workload and system performance. Previous analytical tools have considered either operator workload alone or system performance alone, neglecting the important interaction. In contrast, CREWCUT integrates workload estimation algorithms into a dynamic human performance simulation framework, such that the performance consequences of excessive workload can be observed and quantified. CREWCUT also provides a mechanism for simulating operator coping or workload management strategies and their effects on system performance. Workload algorithms in CREWCUT are based on a multiple resource theoretical model that includes three workload components: 1) single task demands, 2) cognitive interference, and 3) structural interference. Workload is computed as the demand placed on each mental resource at any given moment in the simulation, given the degree of conflict or interference in resource demands at that moment. When demands exceed threshold, workload management strategies are called into play by CREWCUT and attempt to reduce workload. Various performance penalties are assessed, depending upon the particular strategy selected and its degree of success in normalizing workload. The effects of automation on technology insertion can be simulated in CREWCUT in terms of its effects on operator tasks (sequencing, allocation, time, and error parameters), workload demands, and crewstation displays and controls. The workload and system performance effects of display and control media, as well as the physical layout of the crewstation are considered. Both graphic and tabular simulation outputs are provided, supporting a variety of system performance measures.