Research On Detection And Parameter Estimation Of Polynomial Phase Signal

Polynomial Phase Signals(PPS) exist widely in the areas of radar, countermeasure, etc. PPS is an important signal form for radars to realize high resolution, remote detection. With low probability of intercept, PPS is also a key research interest in the field of countermeasure. Proper processing of PPS can lead to target detection, tracking, high resolution imaging, etc.Detection and estimation are important aspects in PPS research. Although existed achievements can solve the problem of parameter estimation under the circumstances of single-component and high signal to noise ratio(SNR), they do not give adequate research on detection properties and can not be applied to the conditions such as multi-component or low SNR which are very common in practice. Under the support of 973 project, the author did the following work:1) Optimal detection and estimation theory of PPS in additive white Gauss noiseThe optimal detection problem of PPS in additive white Gauss noise is systematically studied. The relations between probability density functions of each test statistic and false alarm probability, detection probability, SNR, etc. are given.2) Local optimal resolution theory of PPS in additive white Gauss noiseIn the environment of additive white Gaussian noise, a local optimal resolution method of PPS is proposed. With the prior knowledge of the existence of some signals near a certain point in parameter space, the proposed method can accomplish the optimal resolution of two PPSs with similar parameters.3) Detection and parameter estimation methods of PPS based on polynomial phase transformationA highly effective estimation method of PPS and a LFM signal parameter estimation method based on signal reconstruction and least square are proposed. The former, with higher estimation accuracy at low SNR, has similar computation complexity as the polynomial phase transformation method. The latter has low computation complexity, which avoids error delivering effect, and the mean square error still reaches Cramer-Rao bound at low SNR. 4) Research on detection and estimation methods of PPS based on phase functionBased on Radon transformation and cubic phase function, a signal detection and estimation method for multi-component cubic polynomial phase signals is proposed. The method, with low SNR threshold, overcomes the fake peak problem of multi-component cubic phase function method.5) Research on digital frequency-band division technology of wide-band signalsA frequency-band division method for wide-band signals based on weighed and overlap-add model is proposed. The method, changing wide-band signals into multi narrow-band signals and still reserving the frequency structure, provides a solution for computation complexity and band width match problems of wide-band multi-signal detection and parameter estimation.