Research On Parameter Estimation For Polynomial Phase

The analysis of time-frequency characteristics and phase modulation parameter estimation of polynomial phase signal(PPS) in the noise are the general problems in such as radar field, sonar system, biomedical application, communication research, seismic signal analysis and so on.The analysis of monocomponent and multicomponent PPS has been extensively developed for several years. These methods, however, can’t fulfill the needs of high-estimation-performance on phase modulation parameter which are most met in engineering practice, and the problem based on mixed-signal interference when arguments are similar can’t be solved through traditional means. Hence, how to heighten the accuracy of the estimated parameters and suppress signal interference has research value in the context of passive radar. For the purpose of settling these problems, the work of thesis is shown below:1. For monocomponent PPS, a product general cubic phase function method(IGC) in view of fourth-order non-linear transformation is brought forward. By using the ability of multiplicative algorithm on suppressing staggered terms and inhibiting the effects of noise, IGC enhance the ability to resist noise under low SNR.2. The spurious peaks of the general cubic phase function method(GCPF) for multicomponent is first identified by theory and simulations and hence the identifiability problem arises. By exploiting the different dependence of a autocorrelation- items and staggered- items on time, the IGC is proposed to deal with the identifiability issue. The deficiencies of parameters matching under IGC for multicomponent is also identified by theory and simulations. Taking these problems, a integrated product non-linear transformation method(IPN)-based multiplicative time-frequency distribution is proposed. By using successively filtering method for signal separation, the IPN solves the problem of parameters matching and obtains parametric values for status of lower SNR.3. A hybrid random cubic phase function, based on the jointly of integrated cubic phase function(ICPF) and the random cubic phase transform function(RCT), is proposed for the identifiability problem about the Interference between signals under RCT for mixed PPS signal and enhance the definition on obtaining parametric values for status of lower SNR.4. Synthesize the above method, comparison chart on time-domain for signal suppression before and after of IPN function in support of the actual measurement data about Frequency Agile Signal from non-cooperation Radar is given explicitly. It confirmed the usefulness of this method in parameter estimation.