Research On Acoustic Localization Fusion Algorithm Based On Network Ammunition

Target passive detection technology is an important means of military investigation.Passive acoustic positioning technology is widely used in various environments to detect the target because it is passive receiver and is not easy to be detected.It is highly applied in various military combat systems.Based on the networked ammunition detection system,this paper studies the algorithm of acoustic source target localization in passive acoustic detection system.In this paper,the acoustic target positioning and orientation technology are studied,and the spatial five-element stereo array is established as the basic formation of positioning by analyzing and emulation.Based on the array structure,the passive acoustic location algorithm based on time delay estimation and the high resolution spectral estimation algorithm are studied.In this paper,the orientation of single array location and multi-array data fusion is studied.In the time delay estimation algorithm,the commonly used time delay estimation algorithm is summarized,and the theoretical analysis demonstrates their respective characteristics and selects the most appropriate algorithm in the current application conditions.The error distribution of the two arrays is analyzed.In this paper,a time delay localization algorithm based on BP neural network is proposed to effectively eliminate the blind area of traditional algorithm and improve the robustness of the algorithm.Considering the characteristics of high resolution spectrum estimation algorithms and the principle and the difference of time delay estimation method,established the mathematical model for spatial arbitrary array MUSIC algorithm,five yuan to make three-dimensional array of sound source direction Angle and pitching Angle of 2d joint estimation algorithm simulation.Compared with the time delay estimation algorithm.Several data fusion methods are proposed in the multi-array fusion algorithm.The monte carlo simulation of positioning root mean square error under different fusion conditions is designed and the error model is obtained.A weighted fusion method based on density clustering is proposed under the condition of the number of arrays,and the simulation is made.A solution is proposed considering the delay calibration in the fusion algorithm.