A target recognition system by an intelligent radar

ABSTRACT PhD Thesis A TARGET RECOGNITION SYSTEM BY AN INTELLIGENT RADAR Engin AVCI Firat University Graduate School of Natural and Applied Sciences Department of Electrical - Electronics Engineering 2005, Page: 145 In today's technology, intelligent radar target recognition systems have gained a great importance in various areas. In this thesis, a powerful feature extraction is aimed in order to classify- radar targets correctly. Thus, we have used pattern recognition methods. Pattern recognition is the linguistic definitions which are extracted by the human being from the various sounds, images and likewise patterns. Pattern recognition consists of two stages. These are the feature extraction stage and the classification stage. The most important aspect of the classifier design is the selection of optimum features which will characterize the system. In advanced pattern recognition techniques a better feature extraction becomes a matter of primary importance instead of designing a complex classifier. With this aim integrated feature extraction methodologies were used based on wavelet transform, entropy, and various time-frequency spectral analysis methods for radar target echo signals. By placing the feature extraction methods into the artificial neural network and adaptive network based fuzzy inference system pattern classifiers, adaptive pattern recognition approaches were presented. Besides in this study, feature extraction methods based on optimum entropy approach and various image segmentation were developed by using genetic algorithms. The efficiency and reliability of these methods were examined by using real pulsed radar target echo signals from the Lab - Volt radar experimental set. The success of the system was tested with the noisy echo signals, which were obtained from targets. Correct recognition about 95% was achieved. In various fields such as telecommunication and biomedical, these methods will increase the reliability of automatic intelligent recognition systems by helping the decision maker. Keywords: Radar target recognition system, Pattern recognition, feature extraction, adaptive feature extraction, wavelet transform, time - frequency representations, radar target echo signals, artificial neural network, adaptive network based fuzzy inference system, entropy, image segmentation. XVI