Different Variations of Amplitude Spectrum Based Method for Radar Target Recognition

This study develops a low decision time and highly accurate target recognition method by using wideband scattered signals from radar targets in resonance region. In the mentioned method, estimation of signal parameters by rotational invariance techniques (ESPRIT) is used as main signal processing tool. Several reference ESPRIT vectors acquired by using amplitude based approach, and are reduced to one vector for each target with averaging. These single vectors are used as feature vectors of the target set. The test phase is employed with highest correlation coefficients between amplitude vectors of test signals and feature vectors. The proposed method with two different variations is implemented to small-scale airplane targets modeled with wires. In addition to low runtime, high recognition rates especially for the variation using real frequencies are found at low SNR levels.