A Route Selection Problem Applied to Auto-Piloted Aircraft Tugs

The antithetical needs of increasing the air traffic, reducing the air pollutant and noise emissions, jointly with the prominent problem of airport congestion spur to radically innovate the entire ground operation system and airport management. In this scenario, an alternative autonomous system for engine-off taxiing (dispatch towing) attracts the interest of the civil aviation world. Even though structural and regulatory limitations undermine the employment of the already existing push-back tractors to this purpose, they remain the main candidates to accomplish the mission. New technologies are already under investigation to optimize towbarless tractor joints, so as to respond to the structure safety requirements. However, regulation limitations will soon be an issue. In this paper, a software solution for a route selection problem in a discretized airport environment is presented, in the believe that a full-authority control system, including tractors' selection logic, path planning and mission event sequencing algorithms will possibly meet the regulation requirements. Four different algorithms are implemented and compared: two Hopfield-type neural networks and two algorithms based on graph theory. They compute the shortest path, accounting for restricted airport areas, preferential directions and dynamic obstacles. The computed route checkpoints serve as a reference for the tractor autopilot. Two different missions are analyzed, concerning the towing of departing and arriving aircraft respectively. A single mission consists of three different events, called phases: Phase 1 goes from the actual tractor position (eventually the parking zone) to the selected aircraft (parked or just landed); Phase 2 is the actual engine-off taxi towing; Phase 3 is the tractor return to its own parking zone. Both missions are simulated and results are reported and compared.