Research On Target Echo Detection Technology Of Active Sonar For Small And Moving Platform

Target echo detection is an important aspect in active sonar.Especially for those mounted on the small and moving underwater platform,they usually face some special problems in application.The purpose of this paper is to study the target echo detection technology on the small and moving platform.Although with the advantages of small scale and flexibility,these platforms are usually confronted with special problems in target echo detection.In the light of these incurred problems,this paper focuses on the following four key technologies: sonar waveform design and optimization technology,space-time adaptive detection requiring small training data under compound-Gaussian reverberation,space-time filtering for target echo enhancement under reverberation disturbance,mismatched filtering for range sidelobe suppression.It aims to lay the foundation for the active detection of underwater targets based on the unmanned underwater vehicle.The main research contents and innovations of this paper are as follows:1.For active sonar on the small underwater platform,its transmission capability is limited by the platform scale.Therefore,the selected waveform should present sufficiently low peak-to-average power ratio(PAPR),so as to transmit efficiently and ensure the operating range requirement of detection.In order to relieve the high peak-to-average power ratio of some competitive waveforms,it is necessary to design a waveform optimization technique which tries to maintain inherent waveform characteristics while achieving the desired low peak-to-average power ratio.For the requirement above,this paper proposes an optimization technique to solve the problem,that is “waveform optimization technique with low peak-to-average power ratio via alternating projection”.This algorithm considers the improvement of peak-to-average power ratio as the spectrum shaping problem.Then,based on alternating projection algorithm,spectrum shaping is realized by designing projection constraint sets and corresponding projection operators.Finally,the proposed approach makes a trade off between inherent waveform characteristics and peak-to-average power ratio.Theoretically,its performance involved convergence and effectiveness are explained.Owing to the alternating projection method,the generated waveforms exhibit superior amplitude spectrum approximation and low out-of-band spectrum leakage,which is suitable for the efficient transmission in active sonar.In simulations,the control of peak-to-average power ratio is investigated for comb signal and pseudo-random signal,and they validate the proposed approach.2.The array reverberation on moving active sonar receiver has the coupling characteristics in space-time domains.It cannot be suppressed sufficiently if only by the processing in space or time domain alone,and thus the space-time processing method is required.Besides,the reverberation usually shows non-stationarity and non-Gaussian distribution because of the scale inhomogeneity of bottom scatterer.Hence,a space-time target detection approach capable of suppressing non-stationary and non-Gaussian reverberation is desired.For the requirement above,this paper proposes a STAP-based detection approach,that is “a parametric GLRT method for multichannel adaptive signal detection in space-time correlated compound-gaussian disturbance”.In the proposed approach,the compound-Gaussian(CG)distribution models the non-Gaussian characteristic,and the multi-channel autoregressive parametric model reduces the problem dimension and only requires a small-scale training for adapting non-stationary.Finally,based on GLRT theory,a multi-channel and parametric detector CG-PGLRT is derived.It can also be applied to the degenerate Gaussian background.In simulations and experiments,it can be seen that compared with the traditional detectors based on covariance estimation,the proposed detector has a significantly reduced demand for training data while presenting superior estimation and detection performance.3.Although several anti-reverberation strategies may have been adopted,the filtering for bearing-time-record may still be severely affected by the near receiving reverberation,especially in the shallow sea.Usually,the real target echo will be masked by numerous false reverberation highlights,which will greatly increase the false alarm rate of subsequent detection.Therefore,a method of further suppressing reverberation and enhancing visibility of the target echo is required.Aiming at the problem that residual reverberation affects the subsequent detection of near targets in the bearing-time-record of moving active sonar,this paper proposes several approaches for target echo enhancement based on low-rank and sparse decomposition under reverberation.In the frame-by-frame record,it is assumed that the reverberation component is approximately low-rank,the moving target echo component is sparse,and the residual is the noise component.Thus,a low-rank and sparse representation can be assumed on the data to be separated.Distinguished by data organization,the problem can be expressed as an offline model taking the matrix as the basic input,or an online model taking the vector as the basic input,respectively.For the offline problem,the accelerated proximal gradient(APG)algorithm is applied.For the online problem,the low-rank and sparse decomposition algorithms based on effective reverberation subspace estimation methods are applied.By applying these approaches to target echo enhancement,the simulations and experiments illustrate their validity in low-rank and sparse decomposition.4.In some noise-dominated active sonar scenarios,the control of range sidelobes of correlation is a vital issue.It is not only related to the unambiguous ranging problem,but also involves the masking problem of the main lobe of weak targets in the vicinity of the range sidelobes of stronger targets.Aiming at the problem of high range sidelobes in the matched filter output,this paper proposes a method to suppress range sidelobes,that is “optimal mismatched filter design with a controllable maximum loss and low range sidelobes”.It can flexibly control the maximum loss of signal-to-noise ratio(SNR)and mismatched filter characteristics by weighted functions when achieving the optimality of the sidelobe suppression.Classified by the disturbance to be suppressed,mismatched filters can be designed to minimize the integral sidelobe level(ISL)or peak sidelobe level(PSL),respectively.Then,the second-order cone programming problems are constructed with the sidelobe level as the optimization objective and the upper bound of SNR loss as a constraint.Its optimal solution can be obtained by convex optimization tool.When the Doppler of echo signals is known or unknown,the optimal mismatch filters are designed in the one-dimensional or two-dimensional case,respectively.In simulations,several signals such as discrete phase coded signal,linear frequency modulation signal,pseudo-random signal prove the validity of the design approach.