A model to predict mid‐air and near‐mid‐air collisions

Abstract

AbstractThe probability of mid‐air and near‐mid‐air collisions in terminal areas depends on a variety of factors such as visibility, number of air traffic controllers, communication channels, terrain, mix of aircraft and aircraft speeds, local densities of air traffic in specific subregions surrounding airports, as well as other measures of congestion. We develop, with the use of influence diagrams, a probabilistic model to analyze and predict mid‐air and near‐mid‐air collisions. The probability of near‐mid‐air and mid‐air collisions in the model is influenced by weather, traffic density, detection and collision avoidance maneuvres. Low‐hazard near‐mid‐air collisions (LNMACs) can escalate to critical near‐mid‐air collisions (CNMACs) and eventually to a mid‐air collision (MAC) with the probability of escalation reduced through the use of a collision‐avoidance system. The predictive probability of a collision in the model is the product of three factorsdensity, effectiveness of the collision‐avoidance system, and the conditional probability that independently of the collision‐avoidance system a CNMAC escalates, by chance alone, to a MAC. The paper obtains several conclusions based on sensitivity analysis of the model. Under the assumption that new designs of collision‐avoidance systems will result in much higher detection probabilities than presently available, it is important that both aircraft use the system if there is to be a significant improvement in collision avoidance. For maximum benefit, these collision‐avoidance systems should not reduce the probabilities of avoidance given that a near‐mid‐air collision has been detected.