Study On Fast Tracking Theory Of Dynamic Acoustic Array For Passive Acoustic Target

In the 21st century,the development of high technology produces has far-reaching influenced on modern war environment,many countries pay much attention on increasing the intelligence level during the research and development of weapon and equipment.In this paper,the fast-tracking theory of dynamic acoustic array for single typical armoured acoustic target has been researched based on the application background of Brainpower Anti-Tank(BAT)Submunition.The main research contents contain the characteristics of passive acoustic signal,the observation signal fast preprocessing technology,the feature extraction and separability of passive acoustic target,the data association and trajectory prediction algorithms of acoustic array for ground target tracking,and the fast tracking filter algorithms,Sepecifically,the main research contents are as follows:(1)The characteristics of passive acoustic signal.The acoustic source features of typical acoustic target signal are analyzed based on its producing mechanism and property.Then,the signal interference,diffraction,reflection,refraction,transmission,scattering of acoustic signal in the atmosphere is discussed,and the propagation model and acoustic velocity linearization model are proposed according to the discussion.The acoustic velocity varies with the altitude and temperature and the variation function is presented.Besides,attenuation law of acoustic signal in the atmosphere and Doppler effect is researched so that the tracking environment of moving acoustic array is recognized.Moreover,the blind feature of passive acoustic detecting is point out for the complexity of battlefield acoustic signal and dual acoustic source directionality of passive acoustic detecting.Finally,the chaos characteristic of passive typical acoustic target radiation signal is verified by studying produce mechanism of the target acoustic signal,reconstructing signal phase space and calculating Lyapunov index.(2)The observation signals fast preprocessing technology.The necessity of multisensor consistency data fusion is explain.The traditional data consistency fusion algorithms and their defects are analyized After that,a data consistency method for multiple acoustic sensor based on frequency spectrum similarity algorithm is proposed,and the feasibility and effective of the proposed method is virified by semi-physical simulation test.Further more,after the typical interference signals of battlefield were synthetically analyzed,a method of signal restoration based on adaptive EMD algorithm is given for the denoising problem of single channel signal,and a signal de-noising method for multi-microphones array based on information fusion and improved EEMD is proposed for the denoising problem of multiple channel signals.Finally,the feasibility and effective of the proposed methods are virified by both static and dynamic semi-physical simulation experiments.(3)The feature extraction and separability of passive acoustic target.Firstly,the traditional algorithms of feature extraction for acoustic signals are researched,such as:Zero crossing rate and energy rate,and then an energy feature analysis method based on "cut-off frequency"EEMD(CF-EEMD)is proposed in order to solve the problems of the traditional algorithms.Secondly,the nonlinear features of acoustic signals for typical armored targets are researched based on mathematical morphology and fractal geometry theory,and then,a fractal dimension estimation method of mathematical morphology theory is given and introduced in the chaotic feature extraction of passive acoustic signals.Finally,the separabilities of all the features,zero crossing rate,energy vector feature and the nonlinear features are analyzed quantitatively(4)The data association and trajectory prediction algorithms of acoustic array for ground target tracking.In order to solve the problem of fast data association for target tracking,the methods of data association are researched from two aspects as follows:1)The data association algorithm based on Gaussian linear model.The basic model and the reasonable hypothesis of the target tracking data association for dynamic acoustic array system are set,and then the observation space is regarded as the distance space,so that the accosiation problem can be convert to a class of combinatorial optimization problems.After that,the ant colony algorithm(ACA)is analyzed,and the elitist ant and optimization ranking strategy are introduced into ACA,which forms improved ant colony algorithm(IACA).Then the data association is combined with IACA,and a data association method based on improved ant colony algorithm is proposed.2)The data association algorithm based on Gaussian nonlinear model.On the basic of the proposed IACDA method,a method of target tracking based on IACDA and particle filter is obtained,the MATLAB simulation results show the validity and feasibility of the proposed method.And then,aiming at the problem of track prediction precision for BAT,two grey residual modification models are proposed,which includes grey residual modification model(GRMM)and grey Verhulst residual modification model(G VRMM).Firstly,grey model of target track prediction is established,and the limitations of this model are analyzed,then GRMM and GVRMM are established respectively on line to correct the grey forecast model,Finally,it shows through computer simulation that the method based on real-time residual modification mechanism can reduce the prediction error effectively.(5)The fast tracking filter algorithms.According to the dynamic model of the dynamic acoustic array tracking system,two fast tracking methods of maneuvering target are proposed in order to solve the real-time problem of target tracking for dynamic acoustic array system.1)The fast tracking filter algorithm based on Gaussian nonlinear model.At first,linear filter algorithm and nonlinear filter algorithms are elaborate briefly,such as:Kalman filter,extended Kalman filter,unscented particle filter(UPF).Especially aiming at the sigma selection strategy of the UPF,an interacting multiple model target tracking method based on improved unscented particle filter(IMM-IUPF)is proposed.And the MATLAB simulation results show the validity and superiority of the proposed algorithm.2)The fast tracking filter algorithm based on non-Gaussian nonlinear model.To solve the real-time problem of maneuvering target tracking with missile-borne acoustic array,a fast Interacting Multiple Models-Extended Viterbi(fastIMM-EV)algorithm-based target tracking method is proposed.Firstly,the target tracking model for dynamic acoustic array under the condition of Colored noise.Secondly,a novel algorithm named extended Viterbi interactive multi-model(IMM-EV)is introduced in the target tracking method.In addition,the fixed coefficients filters such as ?-? filter and ?-?-y filter are used to instead of the second and third order Kalman filter in the traditional Interaction Multiple Model(IMM),and then the fastIMM-EV algorithm is obtained.Finally,by comparing to the MATLAB simulation results of the other four different IMM methods,the practicality and high efficiency of the proposed method are verified.