Study On The Methods Of Acoustic Homing Based On Wavelet Transform And Fractional Fourier Transform

Torpedo is a kind of underwater weapon which can automatically search,track,hit and destroy targets and acoustic homing is the key to determining the torpedo attack performance.In this thesis,the problem of signal de-noising,feature extraction of passive target and detection of active target in torpedo acoustic homing have been explored and researched based on wavelet transform and fractional Fourier transform.The main research work of this thesis is as follows:1.Wavelet threshold de-noising method of torpedo acoustic homing signal is studied.After comparing the advantages and disadvantages of the three wavelet de-noising methods,the wavelet thresholding has been chosen.On the basis of the analysis of the principle of wavelet thresholding,the selection of the parameters of the wavelet threshold de-noising method in torpedo acoustic homing has been discussed.Through a large number of simulation experiments,the selection principles of wavelet basis,decomposition level,threshold and threshold function are summarized.2.The method of passive acoustic homing based on DEMON spectrum extraction and wavelet transform has been improved.Based on a thorough analysis of the characteristics of ship radiated noise and the formation mechanism,the wavelet transform is applied to the classic DEMON spectrum extraction.This thesis proposes a method for extracting envelope based on complex analytical wavelet transform,it can select the frequency band which has high modulation amplitude flexibly,estimate the centre frequency and the bandwidth of the modulated signal.The bandwidth of the acoustic homing signal is narrow,in order to make full use of the information of each band,wavelet packet transform is used to divide the signal into a plurality of frequency bands for processing and DEMON spectrum fusion.The processing results of the measured data show that the proposed method is superior to the direct full band processing.3.The wavelet energy feature extraction method of the passive acoustic homing signal has been improved,the problem of wavelet multi-resolution energy fractal feature on the passive acoustic homing signal has been studied.In view of the shift variance of feature extraction using the conventional discrete wavelet transform method,the dual-tree complex wavelet is introduced in the wavelet energy feature extraction of the passive acoustic homing signal,the wavelet energy feature extracted by dual-tree complex wavelet is more stable and basically does not change even the signal is shifted.Based on wavelet multi-resolution analysis and fractal theory,a generalized fractal dimension is proposed.The wavelet multi-resolution analysis energy fractal feature of the measured data is obtained and the results show that the energy fractal feature is effective.4.The detection method of LFM signal of active homing based on fractional Fourier transform has been improved.This thesis makes full use of the advantage of fractional Fourier transform in LFM signal processing,and introduces it into the detection of LFM signal.The simulation results show that the proposed method can detect the correlation peak when the echo signal has a large Doppler frequency offset and the copy correlator can not work.