Research On Complex Blind Separation Algorithm Based On The Signal Kurtosis

Blind Source Separation is one of hotspot in signal processing technique recently, which has a wide application. Independent component analysis algorithm (ICA) is an developed effective blind signal processing technique recently, and it is playing an increasingly impotent role in many areas. With the development of ICA algorithm, it needs to separate complex signal directly or indirectly. More generally speaking, real ICA is a special case of complex ICA. Complex ICA has important theory and application value, and it has a wide application in the speech signal, image processing, antenna arrays and many other areas. In the past ten years, its theory had been developed quickly and a lot of effective algorithms had been made. Currently, complex ICA theory is becoming more mature, the new algorithms are constantly emerging, and ICA algorithm has become a hotspot in international signal processing areas.Firstly, the thesis introduces the basic theory of ICA, the typical ICA algorithms and its performance analysis. The basic theories mainly include information theory, the mathematical model of complex ICA, its solvability analysis and indeterminacy etc. Typical ICA algorithms mainly include maximum entropy algorithm, JADE algorithm, stochastic gradient algorithm and the natural gradient algorithm etc.Secondly, the thesis focused research the cost function based on kurtosis. In complex blind source separation, we often use the signal kurtosis maximization as cost function. In order to solve this problem, the complex standard kurtosis instead of complex kurtosis, and using the complex standard kurtosis maximization as a new cost function, then applying amends the complex original Newton iterative algorithm to optimize cost function. The simulation result of QAM signal separation shows that the improved algorithm has good separation effect; comparison with the separation algorithm of kurtosis maximization as the cost function, the convergence performance of improved algorithm also has improved obviously.Finally, in order to reduce the algorithm complexity and improve the robustness of...