Physics-Based and Parametric Trajectory Prediction Performance Comparison for Traffic Flow Management

Abstract

A series of studies compared the predictive performance of a physics-based trajectory modeler with the conventional parametric prediction system currently employed in the operational Traffic Flow Management (TFM) decision support system. The results indicate that the physics-based system has increased performance over the parametric system for trajectories in which aircraft transition in altitude. These studies include a sample size covering thirtysix 24-hour periods in which traffic from 12 Continental US Air Route Traffic Control Centers were examined. Four TFM metrics were used in the studies: Meter Fix Arrival Time, Departure Center Exit Time, Sector Entry Time, and Sector Occupancy. The charts of the TFM metrics for a majority of the data samples share the same characteristics and strongly lead to a consistent interpretation of the results. These interpretations generalize across the metrics for total sample aggregate (all Centers, all dates).