Target Localization Method Of Vehicle Based On Seismic Signature

Unattended Ground Sensors have seen rapid development in recent years, along with the expansion of their market demands. In military, this system has already been equipped in numerous units for target inspection, identification and battlefield confrontation. In civil application, it is generally used for building protection. However, accurately locating targets through seismic signatures captured by seismic sensors remains a difficulty. Therefore, it is of interest to study the problem of vehicle target localization based on seismic sensor network.Based on the analysis of existing localization methods for ground moving target, the Time Difference of Arrival (TDOA) oriented target localization algorithm of seismic signals is summarized in this paper. Ameliorations are given from the perspectives of time delay estimation, signal propagation speed determination and sensor layout optimization. The ameliorated method is then evaluated through wheel vehicle target localization experiments. Finally, a sensor array based TDOA beamforming approach is established for the target localization problem in unknown geological environment.The basic theory of seismic signal target localization is studied, classical methods are compared. In time delay estimation, the performances of General Cross Correlation (GCC) with different weighted functions as well as the higher spectrum based method are compared in simulation. In location calculation, the pros and cons of Fang’s, Taylor’s and Chan’s algorithms are analyzed. The summarized localization method is then evaluated for seismic signals.Based on the TDOA target localization of seismic signatures, various ameliorations are proposed. In time delay estimation, the Empirical Mode Decomposition (EMD) is used, its application to vehicle signals are studied. In signal propagation speed determination, the quadratic interpolation based method is discussed. In sensor layout optimization, the triangular and linear approaches are compared, an optimized five-sensor cross array layout is proposed, and the calculation process of the ameliorated method is concluded.The actual performance of the ameliorated method for wheel vehicle target localization is testified from field experimentation. The frequency domain characteristics of wheel vehicle seismic signatures are first studied, the working principles, selection and installation of seismic sensors are analyzed, and a detailed experimental scheme is given. From the localization results of heavy trucks and light vehicles, it is demonstrated that the ameliorated method has certain effect in improving the target localization accuracy of wheel vehicles.Focusing on the problem of signal propagation speed uncertainty in anisotropic ground, a sensor array based TDOA beamforming approach is proposed. A target bearing estimation band under four-sensor square array is established in theory along with its maximum bandwidth, thus proving the weak correlation between signal propagation speed and target bearing. The effects of propagation speed and time delay measuring errors to the target bearing angle accuracy are studied through simulation. The results of field experimentation show that the localization accuracy of this approach is higher than commonly used three-axis geophones.The ameliorated method proposed effectively improves the accuracy of wheel vehicle target localization. The sensor array based TDOA beamforming approach established realizes ground moving target localization in anisotropic ground. It can be therefore said that the research conducted in this paper has certain guiding significance and application values for the problem of vehicle target localization.