AIR TRAFFIC CONFLICT MODELS

Matt R. Jardin


Abstract


Semi-empirical random variable models of the expected number of air 
traffic conflicts as a function of air traffic density are derived. 
Model parameters are determined from analysis and simulation of 
real air traffic data. These models are applied to simulated air 
traffic scenarios to analyze conflict properties in various conflict 
resolution strategies. It is shown that under free routing conditions 
(i.e. aircraft do not necessarily fly along structured jet routes),
the expected number of conflicts is well represented by a binomial 
random variable model. Using this model, it is further demonstrated 
how conflict resolution may cause a chain reaction, leading to an 
increased number of conflicts for all aircraft, and how the model 
may be used to predict the airspace capacity for a given conflict 
resolution strategy. In a separate study, it is shown that for an 
iterative horizontal-plane conflict resolution strategy, a random
variable model with the geometric distribution closely matches 
empirical data. This model also predicts the aircraft density at 
which the airspace becomes saturated. It is shown how analysis
of conflicts in the horizontal plane may be scaled and applied to 
the analysis of conflicts in 3-dimensional airspace.