Automation Transparency and Personalized Decision Support: Air Traffic Controller Interaction with a Resolution Advisory System

Abstract

Listen

Translate article

In order to satisfy future air travel demands, there is a need for a more automated and modernized air traffic control. Automation is expected to advance from its current principal utilization as software tools to become an autonomous agent cooperating with the air traffic controller. To facilitate interchangeable, functional and sustainable human-automation collaboration, there is a need to develop better interaction and visualization techniques. Ultimately, automation might be rejected because of the system’s opacity (what is it doing and why?) or mismatch in underlying strategy (I would solve this problem differently). Human-machine cooperation is believed to benefit from automation sensitive and adaptive to individual preferences in problem solving. Furthermore, increased transparency afforded by a decision-aid in regards to its reasoning and problem solving, can positively influence user attitudes including acceptance and trust. In a recent study we hypothesized that both these factors (strategic conformance and interface representation transparency), would influence task performance and willingness to use an automated decision aid in a conflict detection and resolution task. Nine controller trainees participated in two real-time simulations in which they were tasked with directing traffic and maintaining separation in short en-route traffic scenarios. Results showed that controllers perceived and used the two interface representations differently. Even though controllers accepted conformal solutions more often than nonconformal, with a degree similar to what has been observed in previous studies, the effect of strategic conformance was not significant. These findings are discussed in relation to the challenges in diffusion and acceptance of decision-aiding automation.