Carl Westin

7-2 Explaining the acceptance of conformal advisories 159 to the controller, a human-machine interface (HMI) mediates a representation of the airspace environment, its infrastructure (e.g., airports, waypoints, and routes), and its populating traffic. By means of the HMI, the controller perceives and reacts upon conflict cues (grey circles), which reflect the objective parameters, when identifying the conflict situation. Consequently, the way the HMI models and visualizes the world has a great impact on how the controller perceive and act in that world. This explains why controllers’ interaction changed (i.e., more speed and combined head- ing and speed solutions), and agreement and response time measures were reversed, when using the triangle SSD in the Automation transparency study. Similarly to the HMI, the automation rule-base (i.e., algorithms) of the CD&R decision aid identifies the conflict situation by processing the same objective param- eters. Based the conflict cues identified by the automation-rule base, the decision aid provides the controller with a resolution advisory , which is presented in the HMI. Even though the HMI and automation rule-base process the same objective parameters, their translation of them may differ. Furthermore, they may focus on different parameters and assign the different weight and priority, thus describing the mismatch in problem-solving between human and machines. In the conformance studies performed in this thesis, however, the conformal resolution advisories were based on a controller’s own solution to the same conflict. As such, the automation rule-base was replaced by the controller’s problem-solving style as identified by the conformance design based on solution data from the prequel simulation. When a controller is notified of a conflict by the system (or self identifies a conflict), information is searched for the purpose of understanding the conflict (see reversed arrows). The controller develops a picture of the situation and conflict. When controllers encountering the repeated conflict in simulations, they identified and acted upon similar information cues that gave rise to their consistent problem- solving behaviors. Debriefings and questionnaire responses suggest that controllers were unaware of repetitions. The controller then determines, through memory retrieval, whether the conflict and solution is familiar. Cues are gathered and matched with previous experiences stored in memory. Different knowledge is considered, including previous experi- ences, rules, and standard operating procedures such as principles and “no-no’s”. 29 This memory retrieval and information exchange is fast and unstructured. If no , more information is required and the controller reassess the situation and clues. If yes , the conflict can be solved quickly on a skill-based level, highly reactionary. More complex conflict situations require a more heuristically driven conscious eval- uation of the solution (the maybe path). The process can be characterized by the take-the-best fast and frugal heuristic for interpreting information clues in relation to goals and expectations. 20, 268 The