Research On Micro-Doppler Feature Extraction Technology Of Target Rotor

The micro-doppler effect refers to the phenomenon that the radar echo is modulated by the micro-motion of the target and its components.The physical parameters and motion characteristics of the target can be extracted from the micro-doppler echo,so the extraction of micro-doppler characteristics becomes an important research content in the field of radar target recognition.Micro-doppler effect has been widely concerned at home and abroad,but its related technology still needs to be improved.This paper focuses on the rotor micro-doppler feature extraction method based on time-frequency analysis.The main contents are as follows:(1)The micro-doppler echo model of aircraft rotors was established on the basis of deriving the micro-doppler mathematical model of typical micro-motion,and the feasibility of extracting the rotor physical parameters and motion parameters from the micro-doppler echo was verified by simulation analysis.(2)Nonparametric time-frequency analysis and parametric time-frequency analysis are studied.The principle and realization of Short-time Fourier Transform,Wavelet Transform,Chirplet Transform and Generalized Warblet Transform are studied.The advantages and disadvantages of the four methods and their performance in micro-doppler feature extraction of rotors are summarized through simulation.(3)In view of the shortcoming of extreme value in window length of S Transform and Generalized S Transform,an improved S Transform is proposed.The Sigmoid function is introduced into the window function of the improved S Transform,and three window function regulators are added to control the window length within a certain range.At the same time,the flexibility and adaptability of the algorithm are increased,so that the time-frequency analysis can meet the resolution requirements of different characteristic signals.Simulation results verify that the improved S Transform has better time-frequency analysis performance and the effect of micro-doppler feature extraction.(4)Adaptive Multi-synchrosqueezing Modified S Transform is proposed.The idea of multi-synchrosqueezing is combined with the improved S Transform,and the number of times of multi-synchrosqueezing is determined by calculating the Rayleigh entropy.Simulation results verify that this algorithm can estimate more accurate instantaneous frequency and thus obtain higher energy aggregation degree.Therefore,this algorithm has high time-frequency resolution and better performance for micro-doppler feature extraction of rotors.(5)Aiming at the problem that radar echo usually contains noise and clutter in practice,a method of extracting rotor micro-doppler information which can reduce noise and clutter is proposed.Complex Empirical Mode Decomposition and multi-autocorrelation can be used to reduce noise and clutter,and then Adaptive Multi-synchrosqueezing Modified S Transform can be used to obtain clear and high-resolution time-frequency graphs.Simulation results show that this method can extract micro-doppler characteristics of rotors efficiently and with high performance so as to realize aircraft classification and recognition.