Passive Detection Of Non-Gaussian Underwater Targets Based On High Order Statistics

Underwater target detection is the main research hotspot of underwater offensive and defensive battlefields,but with the gradual enhancement of underwater target stealth performance,the performance and effective range of traditional detection methods are decreasing year by year,so new effective detection methods are urgently needed to make up for the shortcomings of current technology.In view of the above problems,this paper systematically introduces high-order statistics,because higher-order statistics are not sensitive to the process of obeying the Gaussian distribution,and the noise received by hydrophones is generally considered to obey the Gaussian distribution,so the introduction of higher-order statistics will improve the performance of beamforming and signal detection algorithms,thereby improving the detection performance of underwater targets.The main work of this article is as follows:(1)Study the influence of the marine environment on the high-order statistics of the sonar receiving signal.Firstly,the basic theory of higher-order statistics is introduced,and the commonly used higher-order statistics are compared and analyzed,and then the target and environment of underwater detection,namely ship radiation noise and marine environment,are modeled,the change trend of high-order statistics of ship radiation noise after being affected by underwater acoustic channels is discussed,and finally the principle of using high-order statistics to improve the performance of underwater target detection is analyzed.(2)Aiming at the problem of poor performance of the Minimum Variance Distortionless Response(MVDR)beamforming algorithm when the target signal obeys a non-Gaussian distribution,this paper introduces the Minimum Dispersion Distortionless Response(MDDR)beamforming algorithm to improve performance when estimating target azimuth and signal boosting.By generalizing the minimum variance criterion of the MVDR algorithm to the minimum dispersion criterion,the fractional order features and higher-order statistics of nonGaussian target signals are used to improve the output signal-to-noise ratio of the MDDR beamforming algorithm.Aiming at the problem that the interference suppression ability of the algorithm decreases when there are unstable interference sources,a secondary constraint zerotrap broadening algorithm with minimized energy output in the region is studied.Finally,the rapid solution of MDDR algorithm is improved based on gradient projection method,and the effectiveness of the algorithm and fast solution is verified by simulation experiments and sea trial data processing.(3)Aiming at the problem of weak target detection with low signal-to-noise ratio after beamforming,this paper proposes a weak target detection algorithm combining blind signal separation denoising and signal detector.Firstly,the optimized variable-step independent component analysis(ICA)algorithm based on polynomial decomposition is improved,which is used to realize the blind separation of single-channel ship radiation signal and noise,so as to improve the output signal-to-noise ratio of the signal,and its operation speed is faster and the separation accuracy is higher.Then,the separated signal output by the algorithm is passed through a signal detector constructed based on the fourth-order cumulant diagonal slice,and a new detector combining variable step ICA and fourth-order cumulative diagonal slice detection is proposed,which comprehensively considers the computational complexity and detection probability,and obtains a detection algorithm with a better detection probability of target signals under low signal-to-noise ratio.