The Study Of Blind Source Separation Algorithm Means On The Basis Information Theory

The 21-century is an epoch of information development rapidly. Information quantity is transferred speedy. So asking the information received is correct and speed. The signal processing became a research hot spot of general interest. Blind source separation is the recovery unknown sources from the observed outputs of an unknown linear systems with static parameters. It is a means of treatment aimed at the signal sources. Not depending on the information of the source signals but the limited knowledge of observed signals, it achieves to separate the source signals. Now, BSS has developed rapidly, and it has been used in many fields such as wireless communication, the procession of array signals, speech separation, image procession, biological medical, seismicdetection, radar and sonar, voice eliminator, and so on.The major contribution of this paper is summarized as follow:(1) Primary theory and research trends of blind source separation algorithm are analyzed , three algorithms of blind source separation based on information theory is expatiated that include the maximum output entropy, the likely estimate, the minimum mutual information.(2) This paper bases on the maximum output entropy and the minimum mutual information of information theory, aims at the hyperbolic function, the exponential function and the arc-tangential function, uses the steepest descent method, derives the new iterative formulas.(3) Using three different source signals (the mixture signals of two sub-Gauss signals and a Guass signal, the mixture signals of a sub-Gauss signal, a Gauss signal and a super-Gauss signal), the simulation with computer shows the algorithms have effective performance.(4) Using five different source signals (the mixture signals offour sub-Gauss signals and a Gauss signal, the mixture signals of sub-Gauss signals, a Gauss signal and a super-Gauss signal), simulating on the computer, comparing their separation affect.