Research On Signal Separation Of Radar Echo In Nonstationary Environment

Nonstationary signal detection and separation is the hot issue in signal processingfield. Small target detection on sea clutter background is a typical application of signaldetection and separation in nonstationary environment. Small target detection on seasurface has important signification in operation on sea, ship navigation and seacondition monitoring. This paper focuses on nonstationary radar echo signal, studiestime-frequency analysis and empirical mode decomposition as methods of nonstationarysignal separation, and uses measured sea clutter data to realize signal detection andseparation. Specific content as follows:Nonstationary signal process method is summarized systematically. The currentcommonly used nonstationary signal process methods are sorted as statistical methodsand nonstatistical methods. Principle and characteristic of these methods are discussed.Time-frequency analysis method of Wigner-Hough transform and fractionalFourier transform are studied, an image feature fusion based method of Wigner-Villedistribution cross-term suppression is introduced and applied to signal separation. In theexperiment which contrasts with two signal separation methods affront, its advantage innoise and cross-term inhibition is verified. Its signal separation performance is tested insea clutter data, and its practicability is proved.Signal separation method based on empirical mode decomposition, ensembleempirical mode decomposition and bidimension empirical mode decomposition andtheir engineering realizing key issues are analyzed. Signal separation effects ofempirical mode decomposition and ensemble empirical mode decompositionexperiments in high sea condition and low sea condition are analyzed. Signal analysisresult of bidimension empirical mode decomposition in two dimensional data oftime-range is discussed.Weak signal detection of sea clutter is studied. Three aspects information of seaclutter data in time, range and doppler are joined up, the sea clutter data cube model isestablished, CFAR based3D signal detection method is proposed to detect signal in seaclutter. The experimental results show that, the method can locate the signal correctlyand estimate signal motion parameters roughly. The measured data experiments provethat, the detection method can effectively detect weak signal on low signal to clutterratio.