Unified methodology for planning optimal four-dimensional flight trajectories at the cruising stage in air traffic management

Abstract

This paper proposes a unified methodology for solving the problems of planning and correcting optimal four-dimensional flight trajectories of an aircraft according to the optimality criteria specified by all participants in the air traffic management (ATM) community, taking into account the influence of wind conditions, no-fly zones, moving zones of difficult weather conditions and other aircrafts. To solve mentioned problems, multilayer neural network models are used to construct userpreferred trajectories based on training the data of completed flights along the corresponding route, the A-star algorithm is used to generate optimal trajectories bypassing stationary and moving zones of dangerous proximity of aircraft, as well as the Bezier curve is used to smooth the piecewise linear trajectories formulated by the A-star algorithm according to the requirements for safe tolerances of aircraft flight trajectories. In order to demonstrate the effectiveness of using the proposed methodology, the authors conducted a series of experiments both when planning optimal trajectories pre-departure and when correcting them in flight, taking into account the presence of danger zones (DZ) in the airspace and without them.