Multisensor object identification on airports using autoassociative neural networks

This paper presents a neural identification method of traffic participants at airports. The traffic objects are observed by several dissimilar sensors. Their information is fused to possible object feature sets. These feature sets are processed by a neural network based identifier, so that traffic participants are identified on the basis of very few observations. The neural network identifier is composed of several RBF networks including networks with autoassociative network architecture.