A new method of radar target recognition based on neural network

This paper provids a new method for recongizing quantity of aircraft in radar targets based on neural network for the first time. The method turns many duplicate cycle intermediate frequency narrow bandwidth signal received by radar which once scans targets into complex envelope signal caculates its autocorrelation matrix, and autocorrelation matrix characteristic value ! make these as character of radar target In order to decrease between characteristic components, highlight divergence,reduce characteristic space dimensions, compress its primary characteristic data through APEX (Adaptive Principal Component Extraction) algorithm use; BP (Back Propagation) algorithm recongnize the compressed data as the quadratic characteristic Rcognition trial shows the aircraft number information extract from intermediate frequency narrow bandwidth signal using this method, influenced little by targets space position or posture and it has stronger curb noise ability To compressing primary characteristic data applying APEX algorithm make targets characteristic stable and efficient, and has a relatively well recognition result in applying BP algorithm to target recogntion.