Fuzzy ARTMAP automatic target recognition from single and multiple wavelength radar range profiles

We investigate the use of the Fuzzy ARTMAP neural network for automatic classification of targets based on their radar range profiles. Tests on synthetic data indicate that arbitrarily high accuracy may be achieved by increasing sufficiently the number of aspects used during training. Improvement of accuracy with a fixed number of training aspects can be achieved by simultaneous use of profiles taken at different wavelengths. The problem of rejecting patterns not present in the training set is also examined