Application of Self-organizing Feature Neural Network for Target Feature Extraction
An effective method, extending Prony algorithm based on Self-organizing feature map neural network, is introduced for radar target feature extraction. The method is a modified classical Prony method based on singular value decomposition and excellent classified capability of Self-organizing feature map neural network which can improved robust for spectrum estimation. Simulation results show that poles and residues of target echo can be extracted effectively using this method, at the same time, random noises are restrained in some degree. It is applicable to target feature extraction such as UWB radar or the other high resolution range radar.
KeywordsMean Square Error Radial Basis Function Neural Network Target Recognition Radar Target Natural Resonance Frequency
Unable to display preview. Download preview PDF.
- 1.Wang, J.: High Resolution Parametric Modeling for Two-dimensional Radar Target Using Prony Algorithm. Journal of Electronics 17(1), 38–45 (2000)Google Scholar
- 2.Nam, S., Kang, S., Park, J.: An Analytic Method for Measuring Accurate Fundamental Frequency Components. IEEE Trans. Power Delivery 17(1), 405–411 (2002)Google Scholar
- 5.Han, H., Chen, Z., Zhuang, Z.: Unconstrained E-Pulse Synthesis Method. Journal of Systems Engineering and Electronics 21(10), 23–26 (2002)Google Scholar
- 6.Guo, G., Zhuang, Z., Chen, Z.: Electromagnetic Feature Extraction and Target Recognition. National University of Defense Technology, Changsha, China (1996)Google Scholar
- 9.Howard, D., Beale, M.: Neural Network Toolbox User’s Guide. The MathWorks, Inc. (2001)Google Scholar