Second Order Moments Based Design And Promotion Of ICA Algorithm

Independent component analysis(ICA) refers to a process that the source signal is recovered only by observed signals and the source signal’ relative independence under the condition that unknown source signal and unknown transmission channel characteristics.The technique of ICA has a widely applications in many field(such as wireless communication,image processing,biomedical signal and mechanical fault detection).In this dissertation,the linear instantaneous mixed independent component analysis algorithms are discussed.This paper focus on the second order moments based BFGS quasi-Newton ICA algorithm and adaptive step-size ICA algorithm,the work included are summarized as follows:1.The paper reviews the development and the research status of the ICA,some specific applications of ICA are introduced.The basic theory and method of the ICA problem is systematically addressed,we give the signal model and analyze the two uncertainties and the three assumptions of the ICA problem,the pre-processing method,contract function and optimized algorithms are summarized.2.The second order moments technique of ICA is introduced,we introduce the adaptive ICA algorithm and the pre-whitening adaptive ICA algorithm.Whitening is a useful pre-processing method in ICA,the correlation between the data and the data can be eliminated by pre-whitening processing of the observed data.Computer simulation results show that the separation effect of the second order moments based pre-whitening adaptive ICA algorithm is better than no pre-whitening adaptive ICA algorithm.3.The contract function is to solve the problem of ICA,which can determine the algorithm statistical performance.According to the contract function proposed by Zhang Weitao etc,this paper proposes a new contrast function,by using BFGS quasi-Newton method of the optimization theory and Householder transform method,a new second order moments based BFGS quasi-Newton ICA algorithm is obtained.Simulation results show that the algorithm can effectively estimate the source signal and has a faster convergence speed and a better stability compared with the second order moments technique based ICA algorithm.4.An important problem in adaptive ICA algorithm is opting step-size.According to separation of independent components in the iterative process of the second order moments adaptive algorithm,this paper presents a new second order moments based adaptive step-size ICA algorithm.The new algorithm can improve speed and reduce the steady-state error,The simulation results show that the algorithm is effective.