Short-term and Long-term Dependency Modeling of Consecutive Range Profiles for Radar Target Recognition

High resolution range profile (HRRP) is being known as one of the most powerful tools for radar target recognition. The main problem with range profile for radar target recognition is its sensitivity to aspect angle. To overcome this problem, consecutive samples of HRRP (or feature vectors extracted from them) were assumed to be independent identically distributed (iid) in small frames of aspect angles in most of the previous works. In this work, an alterna-tive method based on dynamic system is proposed to overcome the limitations of the recent methods such as the independency assumption. Two different feature extraction schemes using spectral features and PCA coefficients are utilized. When the first scheme is employed, modeling is done in short-term, whereas long-term modeling is done while employing the second scheme. Simulation results show superiority of our model over common models us-ing either approach to feature extraction.