Carl Westin

44 First empirical insights though, that such efforts have been hindered by one important limitation - namely, they cannot guarantee similarity between human and machine solutions. We, on the other hand, wanted to ask a more fundamental question: assuming “perfect” au- tomation, which by all appearances performed a high level task in exactly the same way as a given person, would that person accept the advice of such automation? To address these questions, we started with a novel experimental design and simula- tion protocol, which allowed us to capture and replay (in an unrecognizable way) specific ATC scenarios, including controllers’ specific actions. Doing so allowed us to ask the following intriguing question: Do controllers reject their own previous solutions, when they (mistakenly) believe that the solutions come from automation? 3-3 Method 3-3-1 Participants Sixteen professional air traffic controllers from Shannon Area Control Centre (ACC), Ireland, voluntarily participated. There were one female and fifteen males, varying in age between 26 and 44 years (mean = 31). Experience ranged from zero to ten years (mean = 2.5). Twelve controllers were actively working en-route posi- tions, while three were en-route trainees at the end of their on-the-job training. One controller was actively working the tower position. 3-3-2 Simulator The Java-based ATC simulator ran on a portable computer connected to an exter- nal 21” monitor. Interactions was facilitated through mouse and keyboard, allowing participants to control short traffic scenarios. The simulator was based on a proto- type of the Solution Space Diagram (SSD) for ATC, which is currently under de- velopment at the Delft University of Technology. 102 The SSD is a tactical decision support tool that displays color coded “go” (safe) and “no-go” (conflict) regions to facilitate participants’ use of heading and speed resolutions. Conflict regions indi- cate that if the velocity vector is within one such area, there will eventually be a loss of separation (defined in en-route airspace as two aircraft being within a 5 nmi and 1000 feet from each other) and possible collision with the aircraft to whom the conflict region depicts. Appendix B provides a detailed description of the SSD. Feedback from initial simulator screenings highlighted several shortcomings of the prototype ATC SSD. The reference group of controllers suggested a simplified SSD better tailored to their working environment. A modified heading band SSD was developed that, among several improvements, incorporated the following key changes: