A fast radar target identification method using SVD noise reduction technique and pmusic profiles

In this paper, a fast radar target identification method suitable for real-time applications will be presented. The indicated method mainly uses singular value decomposition (SVD) based noise reduction technique and pseudospectrum multiple signal classification (PMUSIC) algorithm. In the stage of method's feature database construction, SVD noise reduction is applied to suitable late-time intervals of preselected scattered signals. Afterwards, PMUSIC profiles (vectors) of these noise reduced signals are obtained and average of these vectors is assigned as feature vector for each target. In test stage, PMUSIC vector belonging to test signal is found likewise and identification is done according to the highest correlation between this vector and feature vectors. The proposed method is applied to small-scale airplanes modeled by wires and despite its short runtime, it gives successful results even in high noise levels. Besides, the accuracy rates of method are improved especially in high noise levels with SVD noise reduction.