Font Size: a A A

Distributed Target Accumulated Detection Under Correlated K-distributed Clutter Background

Posted on:2016-02-01Degree:MasterType:Thesis
Country:ChinaCandidate:Y ZhuFull Text:PDF
GTID:2428330473465654Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
Extended target detection in related / non-Gaussian clutter is a key issue to be resolved of high resolution radar.In this paper,a two-dimensional space-time related K distributed clutter model was applied to solve real-time extended target detection,and its application in high resolution guidance radar was investigated.Theoretical analysis and simulation experiments are used to study the modeling and simulation of clutter,clutter normalized and Gaussian process,the generalized K distributed fitting probability density function of the related K distributed clutter power,the order statistics of related generalized K distributed extraneous variables,design and performance analysis of accumulated sliding window detector and the order statistics-sliding window accumulated detector,and other issues.Research work of this paper including:(1)The two-dimensional space-time K distributed clutter model was commonly used to analysis high resolution radar,which the mechanism of K distributed clutter and statistical properties was analyzed by product model,base d on the spectral density function of a typical sea clutter power unimodal(single-pole)property,the normalized correlation function of clutter was simplified to the product of time and space detachable power function,a fast simulation method based on the space-time distribution of clutter first order AR model two-dimensional correlation K was proposed.Simulation experiments were carried out,and through the comparison of statistical properties of simulation clutter data and the theoretical statistical properties of K distributed clutter,illustrates the effectiveness of the proposed clutter simulation method.(2)To the generalized likelihood ratio test of detecting extended target in K distributed clutter,according to the characteristics of the single peak power spectral density function of the clutter,the whitening process is simplified from high-end processing inverse AR and inverse order processing AR,low computational methods for Gauss processing and pre-whitening was proposed,which were based on theory of normality inverse integral transformation and inverse AR(1)conversion.The correlation function of clutter after normalized and clutter before normalized was derived,the effectiveness of Gauss processing and white processing was simulated and evaluated by Monte Carlo experiments.(3)To the problems of energy accumulation or power accumulation in extended target detection,a fitting method based on the moment matching was proposed,which the power probability density distribution of complex random process between the space-time two-dimensional correlation K distribution and the generalized K-distribution.The relationship of Shape parameters,power parameters and two-dimensional normalized correlation function between generalized K-distribution and space-time two-dimensional complex K distribution was derived,the validity of generalized K distribution fitting method was simulation and evaluation by Monte Carlo experiments,which provide the basis for design and performance analysis of target detector in K distributed clutter.(4)To the target detection problem in related space-time K distribution clutter,sliding window accumulation detector and order statistics-sliding window accumulation detector were proposed on the basis of the maximum and minimum radial length constraint criteria.Based on the theory of generalized K-distribution fitting,moment matching,effective sample size estimation of relevant data and fractional order statistics computing,the approximate probability density distribution expressions for the test statistics under null hypothesis are derived.The relationship between false-alarm probability,CFAR detection threshold and clutter parameters are established.Monte-Carlo test is used to evaluate the precision of the fitted probability density distribution and the false-alarm-rate controlled by given detection threshold.The effectiveness and reasonableness of the proposed methods has clearly been demonstrated by theoretical analyses and simulation experiments,which provide an important reference of extended target detection of high resolution radar detector.
Keywords/Search Tags:High resolution radar, Related K distribution clutter, Sliding window accumulation detector, Order statistics
PDF Full Text Request
Related items