| Comparing with traditional narrow band radar signals, owing to increase by radar band width, wide band radars can achieve high range resolution of targets. So wideband radar signal poses many advantages over narrowband signals of traditional radars, for instance, more accurate measurement of position and movement parameters of targets, more and precise information of targets and environment, imaging for targets and mapping for land and ocean, classing and recognizing reliably better radar targets. As one important aspects of wideband radar signal processing, wide band radar signal target detect is different with narrowband radar signal target detect. With the increasing of radar signal band width, range resolutions are improved. Scattering centers of targets are distributed into more range resolution cells along radar radial direction by wideband radar signals, the range spread target signals or distributed target signals are achieved, which the envelopment signals are high resolution range profile signals. Because target echo energy are leaked into reference cells, traditional target signal detection algorithms of narrowband signals are poor or lost efficiency in wideband radar signal detect application, so the detection algorithms should be given specially for wideband radar signals. Scattering center echo energy distributed into different range resolution cells in different echo signal according target motion, especially in different Doppler resolution cells, which is the difficult for wideband radar range spread target detection.The wideband radars researched in this dissertation are space surveillance radars and ballistic missile early warming radars, which main detect targets are satellites and ballistic mollies. These type targets possessed long range characters so background noise of echo signals could be described as Gaussian white noise. The work of this dissertation can be summarized as the following four aspects:Single pulse echo signal detect of wideband radar range spread targets are researched first. Probability distribution of HRRPs analyzed, than a normal distribution detector is presented. The real and image parts of single echo signals are combined one vector, and skewness and kurtosis are computed, than normal distribution test for the vector are carried. The detect decision can be made by the normal distribution test results. Theory analysis shown the normal distribution test detector is non-dependent on scattering centers distribution information and computer experiments show it achieved better detection performance than energy integrated detector.The frequency domain and time domain characteristics of HRRPs are analyzed in second chapter. Envelope signals of range spread targets could be expressed as spectrum analysis results, the corresponding Fourier transformations are more harmonic wave summation which the frequency and amplitude are depended on positions and scattering intensity of scattering centers. According above analysis, harmonic wave number estimation detector for range spread targets is given. Theory analysis shown the harmonic wave number estimation detector is non-dependent on target scattering center distribution information, computer simulation experiments shown harmonic wave number estimation detector achieve better detection performance than energy detector for the targets which scattering centers distribution is sparseness.The characteristics for correlation coefficient of two conjunct wideband radar range spread target pulse echo signals are analyzed from more scattering centers model, and correlation coefficient summation detector are presented. Probability distribution of correlation coefficient of two Gaussian noise signals without target echo signals is analyzed than constant false alarm rate threshold is given. Wideband radar range spread targets described by frequency slow time subspace model with uniform motion are analyzed. For each range resolution cell the wideband radar echo signals can be expressed as a single frequency complex harmonic wave signal, and Fourier transformation could be used for target echo signal energy integrated. Then slow time spectrum integrate detector for wideband radar range spread target echo signals is presented. Theory analysis and computer simulation experiments with simulating data show that slow time spectrum integrate detector is non-dependent on apriority information of target scattering center distribution and target motion. Computer simulation experiments shown that both the correlation coefficient summation detector and STSI detector achieve better detect performance than NSDD-GLRT detector.The special relationship between maximum values of correlation series for wideband radar range spread target conjunct echo signals and target motion are analyzed. The correlation number and correlation delay time two dimension plane are presented. The maximum values of correlation series are presented different geometry characteristics according to target’s different motions. Two range spread target detectors, Hough transformation detector and recursive least squares dynamic programming detector are given. Hough transformation detector can detect uniformly accelerative motion targets by detect straight line. Recursive least squares dynamic programming detector can detect uniformly jerk motion targets which achieved quadratic curve lines. The probability for correlation series of two Gaussian noise signals which are computed by discrete Fourier transformation, then the computation method of constant false alarm rate threshold are given. Theory analysis and computer simulation experiments with simulation data show the Hough transformation detector and recursive least squares dynamic programming detector are bath non-dependent on target scattering centers distribution apriority information. Computer simulation experiment results with simulation data shown that both Hough Transformation detector and Recursive least squares dynamic programming detector achieved better detect performance than NSDDGLRT detector.
|
PDF Full Text Request