Dynamic Graph Propagation for Performance-based Tactical Conflict Resolution in Urban Air Mobility

Tactical conflict management is a crucial issue for time-sensitive urban air mobility (UAM) operations, considering safety, security, and efficiency factors. To achieve real-time conflict resolution in structural UAM corridors, the operational environment is formulated as the graph structure, in which the edge connection is the available routes, and the node feature is collected from the flight states, e.g. arrival time, speed, arrival probability affected by uncertainties, and priority. To resolve the short-term conflict, the graph propagation solution is proposed to generate multiple augmented subgraph views based on the prescribed graph, where each subgraph represents one candidate action, e.g. speed adjustment, local re-routing. Information in each subgraph is then aggregated and assessed by the global cost metric. As the consequence, the final action is determined by ranking the cost values of all possible subgraph views. The study cases about the higher-priority intruder and non-cooperative intruder demonstrate the effectiveness of the proposed solution for eliminating the conflicts and reducing the additional cost.