Detection And Parameter Estimation Of Polynomial Phase Signal In Low SNR

The polynomial phase signal(PPS) has been widely concerned as an important signal model in the field of radar, sonar, wireless communication and earthquake. The detection and parameter estimation of PPS is the important aspect of the study of polynomial phase signal. The existing results can better solve the problem of parameter estimation under the premise of the single component and high SNR. However, it is not ideal for application to multi-component and low SNR. So this paper studies parameter estimation of polynomial phase signal for low SNR and processing method of the cross term interference of multi-component PPS. The main work is as follows:① Introduce the basic theory of PPS and describes the two order、three order and high order polynomial phase signal models and presents several typical PPS parameter estimation algorithms. Such as the fractional Fourier transform, short-time Fourier transform, Wigner transform, high order ambiguity function and the three-phase function method.② High order phase function has good estimation performance for the phase signal of higher order polynomial. Because of the introduction of the nonlinear operation, the SNR threshold of the SNR is increased. In this paper, the kernel function is improved to achieve the purpose of reducing the nonlinear operation and reducing the SNR threshold. The algorithm transforms the high order polynomial phase signal into the instantaneous frequency modulated signal by the nonlinear transformation and is easy to estimate. Compare to the high order ambiguity function(HAF) and polynomial Wigner distribution(PWVD). This parameter estimation method of polynomial signals can provide the lower ratio threshold of signal to noise.③ Put forward a phase separation parameter method aim at the existing compression cross technology cannot solve the cross term interference problems of the multi component PPS when high order coefficient are the same. Phase parameter separation method can construct symmetric cross term and eliminate the cross term interference using symmetry. Signal is separated into two signals. The phase parameters of PPS will be divided into two sets and each item has half of its phase parameters. The separated signals were determined by nonlinear processing. Estimate the two groups of parameters respectively by searching the peak value of the spectrum. Compare with high order ambiguity function algorithm, this method eliminate the cross term interference of multi components, reduce the search complexity and improve the parameter estimation performance under low SNR.