Study On Separation Of Radiation Interference Based On Blind Source Separation For Marine Electronic Equipment

Because of the stringent requirements of radiation limit,the one-time pass rate of the radiation test is low in marine electronic equipment,so the rectification is indispensable.However,because of the large number of interference sources,lacking scientific and effective searching method,most equipments have to be adjusted repeatedly to reach the standard.The rectification period is long and it's demanding for the rectification engineers.In order to solve this problem,the paper studies the separation method of radiation interference which is based on blind source separation.The radiation data is processed by the algorithm to obtain the potential information of interference source from radiation data,which can make contribution to finding the reason about radiation exceeding quickly.The details are as follows:(1)Study the mechanism of radiation interference;due to the shortcoming of traditional searching method,propose the need for radiation interference separation;design radiation interference separation scheme based on blind source separation.(2)Specify the test system and method for the data in interference source separation scheme,and study date preprocessing technology.(3)Study the separation of radiation interference signal under positive definite conditions.First,the paper analyses principle of FastICA;then transforms the radiation emission separation problem to the basic problem of blind source separation,and analyzing the feasibility;finally,separates the radiation interference signals in simulation and testing.(4)Study the separation of radiation interference signals under over-determined condition.First,the paper analyzes principle of the SCA briefly;then,due to the shortcomings that the number of source signals are necessary in advance,uses source number estimation method which is made of ensemble empirical mode decomposition and singular value decomposition;finally,separates the interference signals by source number estimation method and SCA algorithm.