Self-Organizing Map (SOM) Neural Networks for Air Space Sectoring

Abstract

Self-organizing map (SOM), an unsupervised learning way of artificial neural network, plays a very important role for classification and clustering of inputs. The property of SOM, also called topology-preserving maps or self-organizing feature map (SOFM), is observed in human brain which is not found in other artificial neural networks. Aircrafts' crossing points between two airports may generate conflicts when their trajectories converge on it at the same time and induce a risk of collision. This risk of collision can be avoided by using the self organizing map neural network clustering algorithm. This paper presents the computation of automatic balanced sectoring of airspace to decrease collision and increase air traffic control capacity in high density traffic airspace area. Moreover, SOM is found better technique in comparison to the ART1 neural networks a genetic algorithm used earlier for the same problem.