Matched Field Localization In Waveguide By Three-dimensional Helix Array Fourier Transform

Passive source localization is one of core issues in underwater acoustic detection.Source localization in nature,is an inverse problem,that gives a radiating source location's estimation based on analysis of spatial-temperal records and known a prior knowledge(propagation models and environmental parameters).Underwater acoustic waves propagate in an upper-and lower-bounded waveguide.This indicates the intrinsic three-dimensional property of source location and acoustic field.A horizontal or a vertical linear array,which is commonly used to collect one-dimensional sample of acoustic field,determines sources' bearings in angular space or positions in depth-range plane.A linear array is incapable to determine sources' three-dimensinal coordinates,because of the absence of three-dimensional sampling.This dissertation solves it by studying a three-dimensional array,helix array,with three-dimensional joint localization methods.It's obivious that a target of interest appears on a single point in the space,meaning its existence function is a Dirac delta function.This requires estimated ambiguity function to be a better approximation of the delta function.Source localization in waveguide mainly uses matching results between observed data and simulated data generated by propagation models.The maximum point on ambiguity surface corresponds to estimated source's location.A matching procedure can be done in array data domain or in wavenumber domain.Wavenumber spectrum is related to array data by Fourier transform,but array data domain methods are much more adopted.In realistic scenario,array sensors only collect finite samples on space and time domains.According to Gabor's uncertainty principle,the discrete finite samples may cause biase and leakage in its corresponding spectra domain.This will introduce sidelobes on ambiguity surface.A weighting function is needed to concentrate energy,narrow mainlobe width,lower sidelobe level,and finally approach Dirac delta function.The inverse of channel response is the optimum one.However,it's not always available.Then,the adjoint of channel response is a more general and robust one.It's realization in time domain is time reverse,spectra domain is conjugate.Both lead to matching filter.This dissertation uses three-dimensional Fourier transform and does matched processing in wavenumber domain to approximate delta function and achieve under water source localization,from concepts of the space theory in mathematics.This method can also be treated as a spectra estimation method.This dissertation starts from spatial Fourier transformation and the characteristics of helix array's spatial Fourier transformation.Also,its degradation cases,circular linear array and vertical linear array,are considered.Simulation results of matched field localization by three-dimensional helix array Fourier transformation are shown,in comparison with the conventional matched field processing.A waveguide water tank experiment is presented in final part,validating the method proposed here.