Study On Target Detection And Imaging Methodologies For A Widely--Spaced MIMO Sonar System

Detection,localization,tracking,and identification are four important topics in the underwa-ter acoustic detection system.The performance of an active sonar system is not only affected by the fluctuation of the target's scattering strength,and spatial-temporal-frequency variation of the underwater acoustic channel just like passive sonar systems,but also limited by active transmis-sion resulting in reverberation.Those influences degrade the detection performance of conventional monostatic or bistatic sonar systems in shallow water.A widely-spaced multiple input and multiple output(MIMO)active sonar exploits spatial diversity to improve performances of target detection and localization by incoherent signal processing and improves the stability of target detection at the same time.Besides,if the receivers and transmitters of a widely-spaced MIMO system are spaced sufficiently,scatterings from different aspects of the target can be obtained for the inverse scattering imaging of the target.Firstly,the research of this dissertation focuses on the signal model of a widely-spaced MIMO active sonar.In active detection,the primary task is to raise the signal-to-reverberation ratio(SRR)and identify the targets' echoes from the reverberation background.Therefore,the research starts from two aspects of effective illumination at the transmit end and high-resolution processing at the receiving end.According to the sonar equation,effectively illuminating a target means more energy should be projected on the target with less energy leakage in the reverberation area.This dissertation focuses on a widely-spaced time-reversal MIMO sonar system and a widely-spaced phased-MIMO sonar system to suppress reverberation and raise SRR through iteration of transmit-receive interaction for high-directivity illuminating the target.For high-resolution identification at the receiving end,a two-dimensional deconvolution method is developed to improve the resolution of the matched filter and conventional wideband beamforming's output in the beam-delay domain when implemented on a uniformly linear array due to narrowing the main lobe and lowering the sidelobe of the ambiguity function.The performance of detection and localization can be further improved when the deconvolution method is combined with the widely spaced MIMO sonar.This dissertation then focuses on the inverse scattering imaging problem after target detection and localization.An underwater target is usually a combination of several streamline objects,which exhibit a specific angle-frequency dependent and geometric size-dependent scattering characteris-tics.In active detection,the object is modeled as a point target with a certain scattering strength which neglects the information of a target's geometric shape carried by scattering waves.By in-vestigating the acoustic scattering models,this dissertation derives the acoustic scattering pressure expression of a vertically-extended finite length rigid cylinder placed in a waveguide.Also,the investigation is carried out how environmental parameters and object types affect the defocusing effect and ghost imaging.After localizing a target and analyzing its wideband scattering waves,the inverse scattering imaging method is applied to invert the sphere or cylinder's shape.Different from the free field scenario,in a waveguide,mode coupling in scattering procedure and multi-mode propagation result in defocusing and ghosting in inverse scattering imaging.Thus,phase-conjugation based- and sparse reconstruction(SR)based-channel deconvolution inverse scattering imaging methods are proposed to solve these problems.The former uses the time-reversal concept to reduce the effect of channel response,while the latter exploits the sparsity of channel response to identify the multi-path components with high-resolution and analyze the distribution of multi-path in the angle-time domain.The SR-based method solves ghosting induced by multipath effect and defocusing introduced by lower group speed of each mode in comparison with sound speed in water.The SR-based method improves the image quality due to suppressing the sidelobes af-ter high-resolution identification of multipath components.Moreover,the SR-based method also helps alleviate the mutual interference and multi-path caused orthogonality degeneration between waveforms in a MIMO sonar.Numerical simulation and experimental results have proven that the proposed widely spaced MIMO active sonar outperforms the bistatic sonar in suppressing reverberation and high-resolution localization.Compared with conventional beamforming,the two-dimensional deconvolution beam-forming method has a higher resolution in the beam-delay domain.The experimental results have verified the effectiveness of the inverse scattering imaging method proposed in this paper after target localization is performed.