Penetration Overload Signal Feature Extraction Based on MOMEDA and Mutual Information
In the process of ammunition penetrating into multilayer buildings at high speed,the penetration o-verload acceleration signal presents the characteristics of adhesion and aliasing,which heavily affects the accu-rate identification and extraction of penetration features,and makes it difficult for the fuze to count the layers of targets accurately.To solve the above problems,this paper proposed a feature extraction method for penetration overload signal based on multi-point optimal minimum entropy deconvolution adjustment(MOMEDA)and mu-tual information.Considering that the response law of the fuze acceleration sensitive system was unknown under high-frequency strong dynamic loads,the proposed method utilized the non-iterative blind deconvolution en-hancement technology of MOMEDA to achieve noise reduction for the original penetration overload signal.To further enhance the highlighting of multi-layer target features in the original overload signal,the length of the MOMEDA filter was further optimized based on the mutual information theory.Finally,the verification results of the simulation and test signals of the fuze indicated that the proposed method could effectively highlight the penetration characteristics in the original overload acceleration signal,which providing a basis for the accurate layer counting function under strong aliasing signals.
万方数据
维普
