An Influence Model of the Human-Automation Team: Effects of Workload and Automation Reliability, Transparency and Degree

Abstract

Four important variables that influence performance of a human-automation team (HAT) are the workload (resource) demands of the task and environment, and the reliability, degree, and transparency of the automation that assists the human. In this paper, first, we present a model that predicts how these variables influence the human contribution to performance of the HAT, as mediated by changes in trust, dependence, and situation awareness (SA) and experienced workload (dependent variables). These factors vary in their strength of influence and their interaction with each other. Then we describe in greater depth how a meta- analysis of 50 studies has revealed differences in the strength of influence of automation transparency on those dependent variables. In particular, transparency has a large effect on improving the accuracy of performance, increasing trust and situation awareness. Transparency has different benefits for performance in routine situations, where it improves accuracy of the human-automation team, than for situations when automation unexpectedly fails, where it decreases the time for failure recovery. Finally, we illustrate how the model accommodates differences in discrimination task difficulty revealed in an experiment on decision aiding for nautical collision avoidance. This shows the benefits of the model in predicting the tradeoff of factors influencing human-automation team performance.