Study On Filtering Method Of Stationary Random Signal Based On Correlative Characteristics

Random signal processing is a significant branch in modern information science and technology, and it is also one of the effective ways to study random phenomena in natural science and social science. Stochastic signal processing technology is signal processing theoretical bases including the target tracking and detection, estimation, filtering and so on, which has a wide range of applications, such as communications, earthquake prediction, weather forecast, statistical physics, operational decision-making, image processing, biomedical as well as other fields. In recent years, with the rapid development of the modern communication technology, modern information processing technology and computer science and technology, the theory of random signal processing will continue in-depth, and its application of the method will become increasingly widespread. Random signal filtering is a crucial step in the random signal processing, and it is also one of the focus in the random signal processing field. In this paper, according to auto-correlation function of random signals, we in-depth study the filtering method of the stationary random signal for the problem of noise (especially the singular value, also known as outliers) in the random signal. First, according to the relationship between the one step difference variance of sequence and one step difference variance of auto-correlation function, we have come to the law of one-step singular correlation for identification singular value, and the implement method and procedures of identification and treatment for singular value are given, and simulating and studying for it through the MATLAB simulation platform. Research and simulation results show that this method is effective for singular value of deviate from main part from random signal, and has a better effect on identification and treatment to singular value from the smaller amplitude and deviate from the relativity of main part. Second, the method is studied and simulated for the filtering divergence problem from the singular value in the process of Kalman filter. The results show that this method can effectively solve the problem of divergence from singular value; to the problem of modeling precision from the random through, studied and simulated using the method of pretreatment are given. The results show that the accuracy of the model through pretreatment of this method is improved. Finally, conclusions and prospects of this article are given.