Research On Detection And Separation Of Weak Radio Mixed Signals Based On Improved Blind Source Separation

In the development process of the radio,the wide use of various electronic products makes the electromagnetic environment increasingly complex.The required radio is usually mixed signals mixed with a lot of noise,and some useful signals are even submerged in the noise,which makes it difficult for people to obtain and utilize them.Therefore,the detection and separation of weak radio mixed signals becomes an important research subject.In recent years,blind source separation algorithm has been widely used in signal processing with its unique advantages.The models of signal mixing in blind source separation algorithm include overdetermined,determined and underdetermined mixed model.However,blind source separation algorithm is more sensitive to noise and is constrained by underdetermined mixed model.According to the problem of detection and separation of weak radio mixed signals,blind source separation algorithm is improved in this thesis,and mainly studies the following three aspects:Firstly,according to the problem of detection and separation of weak radio mixed signals by blind source separation algorithm under overdetermined and determined mixed model,the method is proposed by using time domain accumulation method to detect signals under this condition,then using wavelet packet analysis for further processing,and then using RobustICA algorithm from blind source separation algorithm for separation.In the simulation research and experimental analysis,the before and after detection and separation of results are qualitatively and quantitatively analyzed by observing waveform in time domain figure,frequency in frequency domain figure,similarity coefficient and signal to noise ratio.Both simulation research and experimental analysis show that the proposed method is suitable for the detection and separation of weak radio mixed signals under overdetermined and determined mixed model.Secondly,according to the problem of detection and separation of weak radio mixed signals by blind source separation algorithm under underdetermined mixture model,this paper focuses on the problem of blind source separation under single channel mixed model in underdetermined mixed model.The method is proposed by using the time domain accumulation method to detect signals under this condition,then using variational mode decomposition(VMD)algorithm to coarsely separate,and then using VMD algorithm withoptimized parameters to coarsely separate.Parameters in VMD algorithm are optimized by using improved quantum particle swarm optimization(QPSO)algorithm.The evaluation indicators of simulation research and experimental analysis are consistent with those of weak radio mixed signals in overdetermined and determined mixed model.And it is proved that the proposed method is suitable for the detection and separation of weak radio mixed signal in single channel mixed model.Finally,Matlab GUI is used as the development platform.The software development of weak radio mixed signals detection and separation system based on improved blind source separation is developed in Matlab.According to the detection and separation method of weak radio mixed signals which is proposed in this thesis,signal reading interface,signal detection interface and signal separation interface are compiled respectively.