Detection, classification and identification of helicopters out of their acoustic signature by soft-computing-based algorithm

摘要

Helicopters are still a kind of weapon which can not be opposed easly by common anti-plane weapons, as they can hide themselfs beneath houses, trees or hills and even when they got detectible in their attacking phase the time for identification and opposing is more or less a few seconds only. For that reason it is necessary to locate, to classify and to iendtify them as soon as possible by non optical classification systems as for example by their acustical,, finger-print``. In this paper we present a soft-computing-based algorithm which can detect helocpters out of a abitary acoustic scene and can identify them by their non-changeable acustical nosie behaviour. We will show that a special preproecessing algorithm based on the so called GAS-and FD-representions enables this algorithm to demodulate acoustic scenes in that way that all kind of helicopter-main-tail-and helicopter-back-tail=frequencies can be detected very clearly and that he integrated neural classifier can use this helicopter-typical features for an identification of 99, 2