| With the development of science and technology, the modern radar which transmits wideband signal has become true. The target echo received by radar whose bandwidth is more than one hundred MHz or several hundred MHz has sub-meter resolution ratio, which is called high-resolution range profile (HRRP). Since the more detailed structure information of a target, which can be used for reliable classification or recognition, is provided by its HRRP data, the research on radar HRRP target recognition has received intensive attention from the radar automatic target recognition (RATR) community in recent years. Feature extraction and classification methods are two main work in radar HRRP target recognition. Many recognition algorithms have been developed for application. Based on the theory achievements, it becomes more and more important to manage these new algorithms efficiently and develop a corresponding demonstration system for evaluating and exploiting these algorithms in the real application.Firstly, the scattering center model and preprocessing method are introduced in this paper, where the methods of feature extraction, classification methods and rejection evaluation are discussed respectively. Then several radar HRRP statistical recognition methods based on subspace models are introduced. Here a frame segmentation method for radar HRRPs based on subspace model is proposed, which can obtain higher recognition accuracy compared with the traditional equal interval partition method. Finally the demonstration system for radar automatic target recognition(RATR) based on HRRPs including the software and hardware exploiting is introduced. The interface of demonstration system is designed with its functions developed by MATLAB7.5 and Visual C++ hybrid programming. The ADSP-TS101 general signal processing board platform is used to complete the target recognition algorithms. The computation burden of each recognition algorithm on the general signal processing board platform is evaluated. |
PDF Full Text Request