Research On Target Number Determination For Conventional Radar

In modern wars, air targets usually attack in saturate formation, which brings about a serious problem for conventional low-resolution radars. Therefore, efficient methods for formation targets number determination is of great military significance for timely analysis of enemy's situation, rapid and accurate reaction in battles, as well as the improvement of the effectiveness of defense systems.This thesis mainly discusses the methods for radar targets number determination based on time-frequency distribution and time-varying autoregressive model. Firstly, the Doppler characteristic and the mathematical model in time domain of radar echoes are analyzed respectively. Secondly, the anti-noisy ability and the time-frequency resolution with short data of the conventional time-frequency distribution based methods are compared. On the foundation of this, two methods for target number determination based on Hough transform and Fractional Fourier transform are proposed, which overcome the intercross interference of the Wigner-Ville distribution. Finally, a time-varying autoregressive model based formation targets detection method is studied in this thesis, based on which a novel method for order selection of time-varying autoregressive model with short data for LMF signals is proposed. Then, with the fusion process of time-varying frequency estimation in different range units, the algorithm is enhanced in robustness and is able to detect multi-formation targets with short data. The experimental results with simulated data and real target echoes verify the correctness and validity of the proposed methods.