Rapid extraction of parameters of Micro-Motion target based on time-frequency analysis and random Hough transformation

Qin'gen Chen,Xinggan Zhang

DOI: https://doi.org/10.13232/j.cnki.jnju.2017.06.020

2017-01-01

Abstract:The distinction of warhead and decoys is one of the main technological bottlenecks in ballistic missile defense system,which determines the success or failure of the anti-missile system to some extent.Micro-motion is commonly referred to small mechanical vibrations or rotation in addition to mass translation which may induce additional frequency modulation on the received target echoed signal,called micro-Doppler effect.In academia,the characteristic of micro-motion is considered as one of the most effective features in ballistic midcourse target recognition.The extraction of the micro-motion characteristics offers new measures for target recognition in ballistic missile defense,and it plays an important role in target recognition.This paper addresses the relationship between parameters of the micro-motion of micro-target and their micro-Doppler features.Based on the characterization of micro-Doppler signal of rotation target featuring sinusoidal frequency modulation signal with three parameters,we proposed a rapid extraction algorithm with small amount of calculation for sinusoidal frequency modulation signal parameters.Firstly,the time-frequency analysis was used to deal with the signal of the target.Secondly,the time-frequency ridge extraction was utilized to get the curve of the frequency conversion.Finally,the random Hough transformation of sinusoidal curve was applied to evaluate the time-frequency curve which is gotten by time-frequency ridge.When the signal to noise ratio(SNR)is small,the estimation precision of the method given in this paper is better than Time Frequency Distribution Hough transformation(TFD-Hough)and when SNR is relatively large,the estimation accuracy of two methods is equivalent.Results of the simulation experiments indicated that compared with TTFD-Hough,the method given in this paper has a greater advantage in calculated amount.When the SNR is 0,the computation time of the method is about 19.9% of TFD-Hough and when the SNR is -10 dB,the computation time of the method is only 11.2% of TFD-Hough.