Download or read book Effects of Automation Predictability and Understandability on Trust in Human-machine Teaming written by Jeremy Friedman and published by . This book was released on 2018 with total page 135 pages. Available in PDF, EPUB and Kindle. Book excerpt: In Human Machine Teaming (HMT) systems the automation must serve as an autonomous teammate with the ability to learn and improve. While future learning capabilities will be designed to improve the performance of the machine, they will also cause the machine's behaviors to change over time in ways that may be unpredictable or hard to understand to a human. These changes may, in turn, negatively impact the human partner's ability to develop and maintain trust with the machine, even if the changes result in an overall improvement in the automation's functioning. This experiment was designed to examine how a machine's predictability and understandability affect trust in the automation. It made use of a simulated search and rescue environment in which a human operator was instructed to team with a simulated Unmanned Aerial System (UAS) to survey a group of damaged buildings for viable entry points. There were two independent variables, the UAS automation's predictability, and the UAS automation's understandability. Predictability reflects how predictable the success/failures of the automation were in accomplishing the goal of correctly identifying entry points. Understandability reflects how easily the operator could understand the automation's methods. The operator could choose to trust the automation to find all the potential entry points on its own or check buildings manually to directly verify the automation's entry point suggestions. The operator was also responsible for answering a constant stream of unrelated questions as a secondary task. Trust was measured both before and after the experiment using subjective ratings. The results showed that decreasing UAS predictability had a significant negative impact on participants' trust in the system. The results also showed that participants were more informed by assumptions about the automation (automation bias) then by the UAS's actual understandability. Regardless of what flight path was observed, participants tended to believe that the UAS was working as expected or intended.