Neural Networks in Identification of Helicopters Using Passive Sensors

An artificial neural network (ANN) based helicopter identification system is proposed. All helicopters emit certain characteristic acoustic signatures, which are specific to them. This is due to some differences in their structures and design. These differences in acoustic signatures can be used with neural networks for detection and classification of different types of helicopters. The conventional system uses the ratio of the main-rotor blade passage frequency (bpf) to the tail-rotor bpf. The ANN is trained to use similar main/tail-rotor information, in addition to the parametric spectral representation technique (reflection coefficients). It is also shown that the classifier performance is acceptable even if data is corrupted with additive noise.