Often,
one operator is in charge of observing and controlling several machines distributed in an environment. Such procedures and workflows can be visually modeled to predict and compare the operator’s task performances for different hardware setups, software and display options.
Operator’s task performance is usually evaluated by systematically studying operators controlling a system prototype in a setup that matches as close as possible the reality. Such an approach promises high-quality evaluation results in terms of data precision, but requires a realistic prototype and subjects that represent the targeted audience. The number of participants, the complexity of the tasks to evaluate and the amount of design alternatives to test is limited by costs and time, which often results in an evaluation performed at the very end of the design, with view subjects, very basic tasks and with a final prototype.
The idea of cognitive modeling is to simulate operators based on psychological and physiological plausible models. These models predict operator behavior and do not require a new setup of a study for each new version of a system. The quality of prediction models depends on the degree of model validity. Model-based predictions, gained by simulation runs can be generated much faster. Therefore, the amount of design variants and the complexity of tasks that can be evaluated are much higher. Additionally, these predictions can be generated for each design cycle as an additional source of information indicating the efficiency of an HMI.