Radar Target Recognition Research based on Wavelet Neural Networ
Liu Wei-ting · Fire Control and Command Control · 2007
The construction of the Wavelet Neural Network(WNN) is studied in this paper.Based on local time-frequency characteristics of wavelet,a method to primarily determine the number of nodes in hidden layer is proposed by deleting those wavelet which do not contain any sample data.WNN constructed by the algorithm in this paper has smaller structure and faster convergence speed.Using the designed WNN as classifier of Radar Target Recognition,good recognition effect is achieved.