To gain an understanding about what information requirements there are for future operators of autonomous drone swarms at airport, this study examined how work at airports is structured today, what staff at airports think about the potential implementation of autonomous drones, and what potential interface components that are necessary to be able to control these drones. Interviews and observations were made at five different airports and air traffic control towers with tower and ground staff. Based on the collected data a Cognitive Work Analysis (CWA) was made to understand the domain and a Thematic Analysis (TA) was made to understand the ideas regarding the drones. Finally, a workshop with fellow researchers was made to generate ideas about solutions and interface requirements, which was analyzed together with the CWA and TA. It was concluded that the main values of the system are to maintain Situational Awareness, Avoid collisions, and Keep time in order to ensure safe flight traffic. A major aspect to make this possible is communication. This was also a main factor for the implementation of autonomous drones, where it was considered important to know what the drones are doing and where they are. However, it was not always necessary for every person at the airport to have complete oversight and control over the drone, but having the possibility to gain that information was central. Having overview of the drones’ present activities, being able to see what they had done, being able to control them directly, and getting notified about relevant information were interface requirements identified. It was concluded that despite there being some constraints and challenges to implementing autonomous drones at airports, there are aspects that can be taken into account and information to be presented in the right way for the future operator to enable implementation of the autonomous drones at airports.

<p>Машинско учење и неуронске мреже су алати који налазе све већу<br />примену у решавању практичних проблема. За креирање неуронске<br />мреже потребан је скуп података, који може бити прикупљен на<br />различите начине. У овој тези је показано да се подаци за тренинг<br />неуронске мреже могу успешно прикупити креирањем веб игре.<br />Сакупљени скуп података садржи стратегије решавања проблема<br />трговачког путника и проблема рутирања возила.</p> / <p>Mašinsko učenje i neuronske mreže su alati koji nalaze sve veću<br />primenu u rešavanju praktičnih problema. Za kreiranje neuronske<br />mreže potreban je skup podataka, koji može biti prikupljen na<br />različite načine. U ovoj tezi je pokazano da se podaci za trening<br />neuronske mreže mogu uspešno prikupiti kreiranjem veb igre.<br />Sakupljeni skup podataka sadrži strategije rešavanja problema<br />trgovačkog putnika i problema rutiranja vozila.</p> / <p>Machine learning and neural networks are the tools that are finding more and<br />more fields of application in solving practical problems. For the creation of<br />the neural networks, data can be successfully collected by creating a web<br />game. The data collected in this manner has strategic solutions for the<br />problems of Travel salesperson problem and vehicle routing problem.</p>