IMPLEMENTATION OF NEURAL NETWORK MODEL ADAPTIVE RESONANCE THEORY FOR AIRCRAFT IDENTIFICATION SYSTEM

• Main Author: ICHSAN UTAMA (NIM: 13202121), NUR
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/12485
• Institution: Institut Teknologi Bandung

Radar will emit electromagnetic waves to identify an object that fly in the sky and to get data related with the object. Usually to make process identification of an object easier, radar will be equipped with interrogator identification friend or foe (IFF) that often called secondary surveillance radar (SSR). Interrogator IFF will send request signal to the object that want to be identified. The object or the plane that equipped with transmitter responder (transponder), after receive request signal from the interrogator IFF, will respond the signal automatically by sent identification code. If the object did not respond the signal that sent by the radar properly, then the object will be identified as black flight or hostile (or at least not friendly). To identify the object that categories as black flight, another way can be used by analyze radar cross section (RCS) data and speed of the object.<p>RCS and the speed of the object that caught by the radar not always same. In order to make identification process can be done quickly and accurate enough needed a system that can identify an object with capability to adapt with changeable input data but also can keep the stability. System that has the kind of capability is a system that implements neural network system. Neural Network that used in this final project is adaptive resonance theory (ART) neural network that can adapt with new inputs but can keep the stability of input that have been learned previously. Experiment result showed, the system that made, can recognize aircraft type based on RCS and speed data.