To Intervene or not to Intervene: The Dilemma of Management by Exception

Abstract:

Future air traffic management architectures propose to give aircraft more flight path autonomy and turn the air traffic controller into a manager of exceptions. This article reports on one experiment in a series of studies that empirically explored the cognitive work underlying management by exception in air traffic control. Active practitioners (controllers, pilots, dispatchers) were prepared on the rules of the envisioned system and presented with a series of future incidents, each of which they were required to jointly resolve. Management by exception turns out to trap human controllers in a double bind, where intervening early seems appealing but is difficult to justify (airspace throughput) and carry out (controller workload problems). Late interventions are just as difficult, since controllers will have to take over in the middle of a potentially challenging or deteriorating situation. Computerised decision support that flags exceptions migrates the decision criterion into a device, creating a threshold crossing that is typically set either too early or too late. This article lays out the intertwined trade-offs and dilemmas for the exception manager, and makes recommendations for cooperative human–machine architectures in future air traffic management.