Study On Doppler Parameter Estimation Of Underwater Acoustic Signal And Its Application

The acoustic source signal emitted by an underwater vehicle consists of several tones superimposed on a broadband radiated noise component.As the tones are usually stronger than the broadband component,they are critical for sonar detection and recognition.Furthermore,the acoustical Doppler effect,which occurs due to the relative motion between underwater tone source and measuring hydrophone,imply the source motion parameters.This dissertation studies both the characteristic analysis and parameter estimation of underwater acoustical Doppler thoroughly.On the basis of Doppler characteristic,the techniques of low-frequency tone noise source localization and passive velocity measurement and ranging are studied intensively.The main contributions are as follows:1.Localization of low-frequency tone noise sources of an underwater vechicle with single hydrophone is studied.To solve the problem that the conventional passing characterstic method is not accurate for the localization of low-frequency tone noise sources,a localization method is proposed based on Doppler shift.First,the Doppler shift feature is accurately extracted by a zoom-spectrm instantaneous frequency estimate algorithm,which combines polynomial chirplet transform with Chirp-Z transform.Then,the closest point of approach(CPA)time is estimated by frequency intersection or least square(LS)algorithm.Finially,the positions of low-frequency tone noise sources on the moving underwater vehicle can be identified by referring to a known source.Numerical simulations and sea experimental results validate the presented method.2.The Doppler shift method needs priori parameters when identifying low-frequency tone noise sources.To solve this problem,a localization method is proposed based on Doppler shift rate.It is found that the curve of Doppler shift rate has a valley value at the CPA time.First,the Doppler shift rate feature is accurately extracted by using the polynomial chirplet transform.Then,the CPA time is estimated by valley search of the Doppler shift rate curve.At last,the positions of low-frequency tone noise sources can be known after referring to a known source.Moreover,a new time-frequency distribution(TFD)called L-phase distribution(LPD)is defined.With the fast calculation algorithm based on S-Method,the Doppler shift rate is directly obtained by LPD.Then,localization of low-frequency tone noise sources can also be realized.Numerical simulations and sea experimental results validate the presented method.3.Localization of low-frequency tone noise sources of an underwater vechicle with a circular array is studied.To solve the overlapping interference caused by sources with same or approximate frequency,a localization method combining array processing and time-frequency analysis is proposed.First,the Doppler signals formed when the tone noise sources pass through the CPA are separated in time domain,by using superdirectivity beamforming of a small aperture circular array.The output signals can be approximated as LFM signals.After the LFM signals from the narrow beam are processed by zoom-spectrm pseudo Wigner-Ville distribution or Chirplet transform,the time-frequency images of the noise signals are obtained.Then,the CPA time of each low-frequency tone noise sources can be estimated by peak search of the time-frequency images.Finally,by converting the time coordinated to space coordinate and referring to a known source,the positions of the low-frequency tone noise sources on the underwater vehicle are identified.As the proposed method uses the joint information of space-time-frequency,it solves the problem that array aperture is limited for identifying low-frequency noise sources by the focused beamforming method.Numerical simulations and sea experimental results validate the proposed method.4.Velocity measurement and ranging of an underwater moving object with single hydrophone is studied.The improvements are proposed according to the two steps in classical scheme.First,following the analysis of inner artifacts of WVD for Doppler signal,MWT is introduced.By matching the instantaneous autocorrelation function with a Taylor extended Fourier revolving factor,the MWT of Doppler signal is able to produce a TFD with satisfied energy concentration and without inner artifacts.Based on MWT,an effective method is given for Doppler shift estimate.Second,we deduce that the WVD of the Doppler signal has a peak at the center of the TFD.Therefore,the CPA time,which is a key parameter in determining the source speed and closest distance,is directly estimated with the Doppler shift.When the CPA time is estimated accurately,Gauss-Newton processing is avoided to minimize the cost function in nolinear LS problem.The effectiveness of the proposed scheme is validated by using underwater acoustic data.