• Login I 
  • Projects

    CARE - Managing Uncertainty

    Uncertainty is inevitable in ATM, and an appropriate perception of uncertainty is essential to good decision-making. In addition, ergonomic and psychological studies show that some uncertainty in the system is essential to maintain user interest, attention, situation awareness and enjoyment in the job. Presenting uncertainty, and asking pilots and controllers to manage it more explicitly, may alleviate some of the feared impacts of increasing automation (de-skilling and loss of situation awareness).

    To date, the ATM system has largely managed uncertainty by seeking to eliminate it. Uncertainty is rarely presented explicitly to pilots and controllers in control rooms, flight briefing rooms and flight decks. This approach may not be sustainable in future as the increasing use of predictive tools, increases in traffic volume and types of aircraft, and the development of new ATM concepts are likely to increase the importance of considering whether and how to present uncertainty.

    The objective of the project was to develop, apply and test the preliminary principles of uncertainty presentation in the context of controller tools which involve Trajectory Prediction (TP). Example tools to be considered are the Tactical Load Smoother (TLS), Medium Term Conflict Detection (MTCD) and the Conflict Resolution Adviser (CORA). DB contributed to the final project outcomes with two key results: (i) the definition of uncertainty from the perspective of controllers, (ii) a video scenario to detail the interaction between the controllers and the Trajectory Prediction tools. On the first point, the key finding was that controllers conceive uncertainty as a reason to act rather than as a characteristic of a situation that needs to be reduced. Uncertainty is a trigger for further action, in order to bring the system in a safer state, i.e. a state where enough slack exists to accommodate for any of the possible outcomes of uncertainty. The consequence is that controllers are not interested in the presentation of uncertainty per se, or in a measure of uncertainty, rather they need to know the type of uncertainty they are confronted with, in order to decide whether this can be addressed by appropriate actions. These points informed the design of the interactions between controllers and tools, as they are pictured in the it can be seen from the video scenario.