Radar HRRP Target Recognition using influence region of samples

The k-nearest neighbour (KNN) rule using Euclidean distance is actually the same as template matching method under the maximum correlation coefficient criterion (MCC-TIMM), which has been widely used in high resolution rang profiles (HRRPs) based radar automatic target recognition (RATR). The nearest neighbor rule treats each training sample equally without consideration of different recognition performances due to its congregation around the other samples coming from the same class and segregation from those of the rest classes. In this paper, we propose an adaptive method that takes into account the effective influence size of each training sample and the statistical confidence with which the label of each training sample can be trusted. The experimental results confirm the effectiveness of the proposed method