A planning tool for predicting en route ATC conflicts and designing ATC sectors

Abstract

The modernization of the Air Traffic Control system has led to a reevaluation of sector performance in order to design a safer and more efficient system for the future. There are, however, few available tools to aid in measuring and optimizing this performance. Existing safety models of air traffic lack the flexibility to represent the broad range of air traffic characteristics. In this thesis, we present a simulation model of en route air traffic which predicts conflict frequencies for a sector under any number of traffic assumptions. We then use the results of our mode! to formulate a sector optimization problem and present algorithms to solve the problem. The model we develop is a stochastic representation of sector traffic. That aircraft is, arrivals and speeds are represented as random processes. While aircraft are assumed to operate under straight line travel restrictions, the route network through which the travel is also stochastic. Thus, deviations in flight paths are easily represented. As our model is a simulation, we make few assumptions regarding the input distributions allowing for a great amount of flexibility. The model has been implemented on the VAXNMS system. While there is a limited amount of available conflict data from the field, initial comparisons show results of the model to agree with actual conflict frequencies. Further analysis of FAA supplied data should yield similar results. - Finally, we extend the usefulness of the model by formulating a center districting problem which uses the predicted conflict frequencies to divide an air center into optimal sectors. The formulation models controller-conflict interactions within the framework of a classic queuing system. With a number of simplifying assumptions, the problem reduces to a quadratic integer optimization and two algorithms are provided to solve such an optimization. Unfortunately, the simplifying assumptions are not necessarily supported by actual data so the usefulness of our presentation is more in providing an example of a mathematical programming approach then in providing a robust planning tool.