Research On Adaptive Filtering Algorithm In Non-gaussian Noise Background

In the process of digital signal processing,It is a common problem which the signal is disturbed by noise,which is caused by the interference of noise in the identification and subsequent analysis and processing and judgment to bring us some trouble.Therefore,it is an important task to detect and extract the useful signal from the mixed signal with noise.At present in the field of signal theory,we used to assume the noise as Gaussian noise,and then according to the characteristics of its distribution function to deal with noise.Our equipment are more complex,in the communication process are encountered non-Gaussian noise.According to the Gaussian noise processing effect is not ideal,the extracted signal is easy to distort or even serious damage and loss,and finally the whole system performance degradation.In this paper,we study and propose a new algorithm based on adaptive filtering theory to eliminate and suppress non-Gaussian noise.First of all,this paper introduces several common noise,and then analyzes the difference between Gaussian noise and non-Gaussian,and puts forward several mathematical models according to the characteristics of non-Gaussian noise.Then,the probability density functions and their statistical characteristics are analyzed.Secondly,according to the principle of adaptive filtering algorithm,we introduce the common adaptive filter and its basic structure and application,then the adaptive filtering algorithm is enumerated.In this paper,the adaptive noise suppression system is a typical application of adaptive filtering.Under the condition of non-Gaussian noise,some common algorithms are not ideal for signal processing,this paper studies a new algorithm.In this paper,the NLMS-Newton algorithm is proposed based on the normalized LMS(NLMS)algorithm and the advantages of LMS-Newton algorithm,and some improvements are given according to its characteristics.After the simulation,we can see that although the algorithm has a large number of operations,but the convergence rate and stability has been significantly improved.Finally,the simulation software MATLAB is used to simulate the new algorithm,and the stability and convergence performance of the algorithm are compared and analyzed.