Improved Taxi Prediction Algorithms for the Surface Management System

Abstract

Airport Surface Traffic Management Systems (STMSs) are emerging as a new technology to assist human operators in the management of surface traffic operations. One aspect of such operations is the routing of traffic. This includes the management of arrival traffic that must land and taxi to gates, as well as the management of departure traffic from pushback to take off. In general, arrivals and departures compete for the same taxi and runway resources. In this paper, we describe the Surface Management System (SMS) architecture and key algorithms that help controllers plan and manage arrival and departure traffic. Furthermore, we present results obtained with a mature SMS system designed for major airports. We then describe several algorithms that we have implemented based on the A* Algorithm and concepts from evolutionary computation as potential enhancements to the algorithms currently used in SMS for surface movement modeling. These algorithms have the potential for improving the accuracy of the taxi transit times predicted by SMS. Emphasis is placed on optimal surface routing given the constraints of limited available resources and the necessity of producing conflict-free surface trajectories. A key factor throughout these studies is the use of dynamically varying costs to model resource utilization.