Research On Multiple Parameters Estimation Of Non-stationary Radar Signals

The analysis methods of nonstationary signals, which are widely exist in our real life, play an important role in modern signal processing. When dealing with the nonstationary signals, the performance of traditional methods in array signal processing may degrade heavily because of the narrowband and stationary signals assumptions. The aim of this dissertation is to use the achievements in modern signal processing, and to combine nonstationary signal processing with array/multiple channel signal processing for joint spatial and time signatures estimation of nonstationary radar signals. This is also an advanced and challenging problem to the array signal processing. We focus on array signal modeling, time-frequency analysis method, parameters estimation of wideband nonstationary radar signals, sources number estimation and array calibration to array processing in the presence of nonstationary signals. The main contributions of this dissertation are concentrating as follows:1. The application of spatial data-adaptive evolutionary spectral in nonstationary radar signal are discussed. A DOA estimation method of nonstationary radar signals is proposed by time-frequency analysis method and Joint-block diagonalization method. An algorithm for improving the angle resolution is proposed based on beamforming technique. A new direction finding method of coherent signals is proposed by constructing Toeplitz matrix using the output of symmetry array.2. The LFM transform and spatial spectral estimation method of wideband arrays model is discussed. A time-frequency parameters estimator of LFM signals is proposed using LFM transform. A DOA estimation technique of wideband array signals is proposed by orthogonality of projected subspaces and Toeplitz diagonalization method. The proposed method can calibrated the array position errors and deal with coherent signals. A frequency invariant beamforming method underling beamspace is proposed to analyze wideband nonstationary array signals.3. A joint estimation technique of time spatial parameters is discussed in undersampling conditions for non-linear frequency modulation signals. A time-frequency estimator is proposed for sine frequency modulation signals using high order cumulating of Wigner-Ville distribution. A new method to estimate the parameters of phase-coded signals is proposed by short-time Fourier transform. This method can distinguish between the binary-phase-coded signals and quadrature-phase-coded signals. And a joint estimation algorithm for the time spatial parameters of the signals is proposed using beamspace ESPRIT method and nonuniform L-arrays.4. The source enumeration of wideband signals underling white/color noise conditions is discussed. A new detection method of nonstationary sources is proposed using frequency domain bootstrap techniques. The proposed method can estimate the sources number under color noise by combining the bootstrap and Gerschgorin disk estimation methods.5. The method of array auto-calibration for wideband sources is discussed. A new method is proposed. This method can calibrate the unknown error of sensor gain and phase, array shape perturb and sensors coupling by assistant sensors.