Radar Target Identification Based On Poles Extracting

With the application of HFSWR in national defense, people focus the investigation of the methods of target identification in high frequency section. The dimension of target in resonance region is similar with the HFSWR wavelength. Because of its invariability, poles is the most effective characteristic signal of the radar target in resonance region. Extracting poles from measured data has been an important problem in signal processing.Theories of linear prediction approach, matrix prediction approach and state space approach are introduced. It can be shown that linear prediction approach is one kind of matrix prediction approach, and the latter is equivalent to the second methods of state space approach, so the Prony method, the KT(SVD-Prony) method, and the MPM method are connected in state space.This paper presents the detailed processes of MPM method. Emulating of MPM method shows its insensibility of noise. Then, some improvement are made on MPM method, preprocessing method can efficient deal with noise and multi-way method can fully extract all poles .At last, a thin-wire target and plane targets are used for poles extracting with MPM method. In processing with thin-wire's data, exact poles are got from every data of nine , it shows the MPM method is effective and target identification using poles is feasible. In processing planes'data, extracted poles are sorted to three kinds and one of them match well from two responses of plane, so it may have some relationships between real poles. To exactly extract poles of plane still need more effort.