Model-based prediction of workload for adaptive associate systems

This article describes a method for predicting future mental states and workload of military helicopter crews, and how adaptive technical assistance is derived. A mission plan and a model of pilot tasks are the basis for predicting future task situations. Combined with knowledge of the mental resource demands of these task situations, workload peaks can be identified before they occur. A task-based, context-rich representation of the crews' mental state enables an adaptive associate system to support the crew, while preventing high-workload task situations. Therefore, the associate system changes the task sharing between the human operator and the automated system online by using different levels of automation and a restrained intervention strategy. This concept is implemented as software agent in a helicopter mission simulator and will be evaluated in pilot-in-the-loop experiments in the near future.