Study Of Air Defense Radar Automatic Target Recognition System Based On Wt And Ann

For low-resolution radar, it can only obtain the up-and-down information of targets, so it’s of difficulties to identify the amounts and types of aircraft. If a kind of effective method that provide the active service low-resolution radar with the function of automatic target recognition can be found, the burden of radar operator will be greatly reduced, and that will be advantageous to the informatization constructions of our country national defense.This paper built models for the echo signals of team planes and the planes of different types firstly, then after analyzing the time domain and frequency domain properties of those echo signals detailed, it proposed an algorithm that using Wigner-Ville Distribution (WVD) and Wavelet Transformation (WT) to process the radar echo signals that is a kind of non-stationary random signal. With this algorithm the amount and type feature of aircrafts can be extracted effectively. Then this paper used BP neural network to classify the radar targets. Finally, for testing the accuracy and feasibility of the algorithm, and intuitively getting the operational process and target recognizing results of radar, a simulation software of radar automatic target recognition system was designed under the Matlab environment. The software was provided with so many functions of scenario making, wave generating, antenna scanning, targets detection, echo display and tracking besides of automatic recognition of plane amount and type. Furthermore the software has certain theoretical research value, especially under the condition of hardware lacking, and it can provide convenience for radar operators training during peacetime. Based on this simulation software as the platform, this paper performed classification and recognition experiments for the team planes and the three types of plane targets, the results showed that the amount identification and the type recognition are of high accuracy.