An intelligent recognition model for dynamic air traffic decision-making

Abstract Air traffic flow management system (ATFMS) is becoming increasingly important due to the rapid growth of air traffic and serious flight delay nowadays. To aware the air traffic flow density and identify the heat airspace in terminal areas of large hub airports is essential for an ATFMS. Due to numerous parameters in air traffic flow, traditional methods based on one single parameter fail to reflect the true complexity relationship between these parameters. This study aims to develop an intelligent air traffic flow heat airspace recognition model using advanced data science technique for establishing a real-time cloud map in the terminal airspace of airports, which attempts to use machine learning models to represent the complex relationship among these parameters. In the proposed intelligent recognition model, high dimensions of parameters (basic parameters, additional parameters and time parameters) are processed to achieve a comprehensive and accurate situation awareness for support dynamic air traffic decision-making. An aircraft trajectories points clustering method is developed to generate a 4D heat airspace map. The basic parameters and time parameters are used to identify the heat airspaces; the changes of additional parameters which influence the heat airspaces are identified and analyzed by use of grid graphs of flight trajectories; probability fitting graphs are used to verify accuracy of 4D results in order to support air traffic decision-making. A case study on Beijing International Airport (PEK) is conducted to test our model and has obtained two main research findings: there are two areas of PEK that have the high density and there are hot peaks at two different heights; flight trajectories and speed of trajectories also effect on the heat airspace. The study realizes that the proposed 4D heat airspace model is better for detailed and accurate information construction, expression of spatial changes, and visualization of multiple parameters of temporal and spatial density and range. It can assist the decisions on airspace allocation, and also have a definite reference meaning on alleviating the contradiction between the current air traffic demand and airspace resource constraints.