*2.5. Simulation Setup and Scenarios*

The simulation was implemented for EDDM airport with its two parallel runway systems. Both runways have an offset of 1500 m and a length of 4000 m. The distance between the runways is 2300 m, sufficient for independent usage of both runways. During the simulation, runways 26R and 26L were in use. The ATCO was in charge of both independent parallel runways guiding the arrivals streams. Departure was integrated into the simulation but handled by the simulator automatically via a departure manager (DMAN). No limitations regarding the aircraft type or weather restrictions were simulated.

In total, seven different scenarios were developed. Table 1 displays an overview of the seven developed scenarios and their composition. In the following, all scenarios are briefly explained. For the human performance, no baseline was simulated during the simulation campaign. Therefore, the results were compared to the reference scenarios (R1 and R2). The reference scenarios, R1 and R2, serving as the baseline for the objective data, were considered important to ensure comparability for the simulation data. The scenarios were based on real air traffic. Data for the reference scenarios where taken from the OpenSky scientific dataset [43]. These data are not validated and may contain inaccuracies. The data for R1 and R2 were composed by ten operating hours selected from October to December, 2021, consisting of 38–40 (R1) and 18–22 (R2) landings per hour from OpenSky datasets. R2, which provides a smaller number of aircraft, was taken into the analyses too, to see the comparison of simulation results for an even smaller number of arrivals. This was done under the presumption that a greater number of arrivals per hour makes it more difficult to enable direct routes due to safety issues. Hence, more aircraft have to fly conventional routes, which leads to a greater value of distance flown per aircraft.

The simulation scenarios were based on a medium traffic load at EDDM, which equals two thirds of the maximum traffic at EDDM [44]. The number of departures was reduced. The traffic mix by aircraft type was based on typical EDDM traffic conditions in 2022 [45]. Since currently the share of aircraft with advanced FMS varies widely among airlines, the amount of 4D-FMS aircraft included in the scenario was used as decisive parameter to distinguish the scenarios. Thus, the four simulation scenarios were developed with different amounts of 4D-FMS aircraft, starting with 20% for a training scenario and increasing up to 80%. Respectively, in the present paper, the scenarios are referred as R1, R2, T, S30 scenario, S60 scenario and S80 scenario; see Table 1. For the simulation scenarios, only the information from flightradar24 on the real callsigns, aircraft types and departure airports was used [46]. The percentage of heavy aircraft varied between 3 and 19%. Each scenario lasted for 45 min. Aircraft were initialised outside the E-TMA area and flew predefined arrival routes towards the boundaries of the E-TMA.


**Table 1.** Reference and simulation scenarios: composition and overview (ARR: arrivals).

<sup>1</sup> Data taken from OpenSky database. <sup>2</sup> Data taken from Flightradar24 [46]. <sup>3</sup> Time interval in which the data were captured.

Each ATCO participated in a full day of simulations. Each day was scheduled into five sections, starting with a briefing and training session (T) to familiarise the ATCOs with the simulation environment. Thereafter, the simulation scenarios S60, S30 and S80 were run. After each session, the ATCOs were asked to fill in the post-run questionnaire (PRQ). Finally, a second run with the S60 scenario was conducted. This run acted as an explorative simulation run with the ATCO to obtain more in-depth feedback about the system from the ATCOs. During the explorative simulation run, individual components (Ghosts, TargetWindows and CSVT) were deactivated and activated one at a time. Participants received some time to test the system when one of the components was deactivated and were asked how this affected their work as an open question. This was followed by detailed questions about each tool. The debriefing took place after the non-explorative simulation runs and was combined with the explorative simulation run. The debriefing questions were modified from the first to the second HITL campaign, based on ATCOs feedback. Finally after a full day of simulation exercises, ATCOs were prompted to fill in a post exercises questionnaire (PEQ). As the order of simulation runs was kept constant for all participants, training effects or effects of exhaustion cannot be entirely ruled out.
