Probabilistic modeling and reasoning of conflict detection effectiveness by tracking systems towards safe urban air mobility operations

This study targets the scenario where centralized tracking systems provide tactical conflict detection for urban air mobility (UAM) flights. In this scenario, the interaction between airspace design and tracking system performances, and its impact on the effectiveness of conflict detection have not been addressed in the literature. To overcome this gap, this study aims at achieving probabilistic modeling and reasoning analysis, to provide references for stakeholders in the design of urban airspace and the deployment of flight tracking systems. A framework integrating multiple probabilistic methods is established. We formulate the event tree of pair-wise aircraft encounters with conflict detection provided by the tracking system. Then Monte Carlo simulation is performed by using an agent-based tool that we develop, to quantify the probability of event occurrences. Finally, Bayesian Networks models are built to infer the dependencies of conflict detection effectiveness on the airspace design and tracking system performances. The outcomes of this study demonstrate improvement in conflict detection that better tracking system performances can provide, and the impact of airspace design under different tracking configurations. The results can be used in the UAM traffic network planning, and systems standardization for UAM flight tracking, towards safe urban air traffic.