Research On Active Conformal Array Sonar Beam Optimization And Reverberation Suppression Technology

Because of its large search angle,low frequency band and space saving,active conformal array sonar homing system is one of the main development of underwater homing system in the future.There are two main problems in the signal processing of active conformal array sonar target detection:the high side-lobe in the beam pattern and the strong reverberation interference related to the transmitted signal.Traditional beam optimization methods included analytical algorithms and adaptive algorithms.Analytical algorithms are limited to uniform arrays.Adaptive algorithms are suitable for any array formation and can produce null in the direction of interference,but have some problems such as misconvergence iteration results and the deteriorated of two dimensional beam direction;For reverberation suppression,Space-time Adaptive Processing(STAP)combines space domain and time domain to suppress reverberation interference in adaptive processing.But STAP is difficult to be applied directly,so it is necessary to study the dimensionality reduction STAP algorithms for large array sonar.Specific studies on the above issues are as follows:1.Beamforming optimizationFirstly,the spatial steering vector of conformal array and the basic model of conformal array beamforming were established.Several typical beamforming optimization methods,such as similar weighting and numerical synthesis were introduced.Secondly,the projection-interpolation windowing algorithm was proposed.By interpolation,the nonuniform array can be projected into uniform linear in the direction of the incoming signal.Array weighted by window function has a beam pattern with lower side-lobe than CBF and similar weighting,lower computational complexity than numerical synthesis.Thirdly,two beam optimization methods based on convex optimization for two dimensional beam optimization were obtained.Lowest side-lobe level under the same main-lobe width constraint can be obtained by the first method.Side-lobe level that meets the set requirements and higher array gain can be obtained by the second method.2.Reverberation suppressionFirstly,the reverberation distribution characteristics,the space-time coupling characteristics of the reverberation of active sonar and the basic principle of STAP were studied.Full-dimension STAP was found had the following problems:high training samples are required to estimate the covariance matrix,the computation of covariance matrix inversion is great.By analyzing the active sonar system’s degree of freedom and reverberation’s degree of freedom,it’s found that space-time dimensionality reduction can be used to suppress reverberation.Secondly,the basic principle of dimensionality reduction STAP was studied.The cascaded time-space adaptive,extended factor approach(EFA)and joint domain localized(JDL)were introduced.An improved EFA method:(IEFA)based on the first dimensionality reduction by EFA was proposed.Compared with EFA,there are less computation and training sample requirements.Compared with JDL,better spatial reverberation processing effect can be obtained by IEFA.