The Research Of Detection And Estimation Of Non-Gaussian Noise And Non-Gaussian Signal

The research of non-Gaussian noise is one of the core contents in the modern signal processing, and the research of non-stationary non-Gaussian noise has become a hot and difficult point at home and abroad in recent years. According to the basic theory and method of classical signal processing, applying modern signal processing methods such as higher order statistics and maximum likelihood function, the theory of detecting the stationary non-Gaussian noise and signal as well as the theory of estimating the probability density function of the non-stationary non-Gaussian noise is studied. This research project is of certain utilitarian value. The main work and originality in this thesis can be summarized as follows:1. In accordance with the shape of the probability density function, four kinds of simplified mathematical model on multi-modal process is defined. Based on higher order statistics and the conception and character of the bispectrum belonged to higher order spectrum, the bispectrum algorithms is improved and employed to give out the method of signals detection in multi-modal noise.2. Using the 3rd order cumulant, the method of stationary non-Gaussian signals detection in Gaussian noise is studied and compared in detail with traditional autocorrelation detection method through simulation.3. Based on the basic theory of maximum likelihood function, the likelihood expressions of the probability density function (PDF) of a non-stationary non-Gaussian noise are reduced, the algorithms of estimating the PDF of non-stationary Rayleigh noise and non-stationary glint noise in radar tracking system. In the proposed algorithm, the PDF of non-stationary non-Gaussian noise is evaluated in two stages: a priori update and a posteriori update.