The Application of Automation Systems for Training – Implications of Trust
(Room S320C)
27 Nov 18
3:00 PM
-
3:30 PM
Interactive automation has the potential to decrease the resources needed for effective aviation training systems. Currently, some training environments require instructors to monitor and train multiple individuals at once, while simultaneously acting as fellow crewmembers in the scenario. This creates high workload for instructors and decreases the efficiency of training and the quality of feedback. Leveraging automation to support instructors can reduce manpower requirements, time demands, as well as enable cost savings.
However, experience with new automated systems (positive or negative) influences instructors’ trust in the technology, and therefore, future use and reliance on the system. For example, the instructor may become over-reliant on the automation and miss important aspects of the training. Conversely, they may underutilize the system, thereby increasing their workload. Trust will also influence trainees’ usage of automation, altering the way each individual trainee interacts with the simulation and impacting the fidelity of the training received. This paper will expand on these potential issues, their effects on instructor and trainee behavior, and the subsequent implications for assessment of training. A protocol for measuring trust in automation within a training session will be suggested and outlined.
Lastly, the authors will provide an overview of a use-case Navy effort to develop synthetic, autonomous agents for P-8A crewmembers. The effort’s technology will allow trainees to interact with simulated crewmembers, enabling instructors to focus on instruction and performance assessment. This paper will detail how potential issues with trust in automation are being addressed within this applied context.