Blind Source Separation And Radar Signal Processing Research

In the last several years, the approach of blind signal processing was developed greatly. From the viewpoint of blind signal estimation, blind signal processing can be divided into two categories: blind source separation and blind deconvolution. Recently many excellent algorithms about blind source separation and blind deconvolution were introduced, especially for one dimensional signals and multichannel one-dimension signals, and these algorithms are widely used in the field of array signal processing, speech signal processing, biomedical signal processing, mobile telecommunication and image processing, etc. Radar signal processing based on blind signal processing was being progressed omnidirectional research ,and acquired some of fruits. In this thesis, we mainly studied blind source separation algorithms, jamming and antijamming of radar signal, blind source separation algorithms for radar signal analysis , and time delay estimation based on general radar signal. The main work of the thesis consist of following parts:1) the methods of blind signal processing are systematically studied. In all kind of methods of blind source separation, the hypothesis of statistically independence between source signals is utilized, Blind source separation methods may be divided into algebraic method based on high-order cumulants and iterative estimation method based on information theories. And blind deconvolution is a extension of blind source separation .A sort of blind source separation algorithm-FastICA is simulated.2) Research on the signal models of radar signals and noise that often were used, jamming and antijamming technology. Theoretic gist and research background were provided to radar signal processing.3) Research on blind source separation algorithms applied on radar signal antijamming processing. The noise removing method based on Gaussian moments for separating multiform signals is proved by simulation that the method effectually separate multi-way radar signals polluted by noise and jamming, antijam gusty pulse and remove noise. Simulations prove the algorithm is simple to be implemented and fast convergence.4)Research on time delay estimation technology for radar signal. Because of the time delay estimation based on generalized cross-correlation(GCC) for general radar signal is not accurate enough ,a new time delay estimation method---GCC based on signal phase information is proposed . Simulation prove this method is sufficient.5) In the summary of the dissertation, the viewpoints for future research are proposed.