Study On Detection And Parameter Estimation Of Complex Modulated Radar Signal With Time-varying Parameter

The radar systems are developing fast and the radar signals are becoming more and more complex.Many complex modulated radar signals have been employed in modern radar systems and the modulation parts of these signals contain time-varying parameters.Compared with the common radar signal,the complex radar signals have better detection and stealth abilities.Thus,the complex radar signals bring a new challenge for radar signal reconnaissance and non-cooperative radar signal processing.Recently,some novel non-cooperative receiving structures have been designed and the Nyquist folding receiver(NYFR)is a typical novel non-cooperative receiver.The NYFR can realize ultra-wideband receiving using a small amount of equipment.As a tradeoff,the NYFR output is more complex than its input and it is a typical complex modulated signal with time-varying parameter.Therefore,in order to process the complex modulated radar signals with time-varying parameters,it is necessary to investigate the corresponding detection and parameter estimation methods.Firstly,this dissertation studies the detection and parameter estimation method of the complex radar signal under the common receiving structure.Then,based on the NYFR structure,the time domain parameters of the complex outputs are analyzed.According to different signal models,this dissertation proposes several detection and parameter estimation approaches for the NYFR complex output.Compared with the existing signal processing methods,the proposed methods have better signal processing performances.The main contributions and innovative results of this dissertation are as follows.1)The detection and parameter estimation method for the linear frequency modulated/binary phase coded(LFM/BPC)signal is proposed.According to the energy detection theory,the detection method is analyzed in the Gaussian white noise environment.The LFM/BPC signal is modeled as a vector and the square processing is used.Then,based on the subspace and pseudo-spectrum theory,the chirp rate estimation method with high accuracy is proposed.According to the chirp rate estimation result,the de-chirp process is used to the LFM/BPC signal and the signal is modeled as a Hankel matrix.Based on the Hankel-singular value decomposition(SVD)algorithm,a phase changing point position estimation method of the binary phase coded(BPC)signal is proposed,which can obtain an accurate estimation result.Combining with relationship between the singular vector and the eigenvector,the singular vector of the Hankel matrix can be calculated and the reason of the sensitivity of the Hankel-SVD algorithm to the BPC modulation part is proved mathematically.2)The NYFR uses a non-uniform local oscillator(LO)to realize the wideband receiving and its output is usually a complex signal with time-varying parameters.According to the typical NYFR structure,the NYFR non-uniform sampling is converted to the uniform sampling based on the extended Fourier transform(EFT)theory.According to the uniform sampling in EFT time domain and the uniform sampling theorem,the time domain characteristics of the NYFR complex output and the effect of the NYFR parameters on the time domain distortion are analyzed under on the low-pass sampling,band-pass sampling and crossing Nyquist zone(NZ)sampling conditions.A trapezoid energy envelope and wavelet transform method is proposed to detect the NYFR pulse output.The sensitivity of the wavelet transform to the trapezoid envelope is proved.This method is applicable to the crossing NZ sampling condition.3)The essence of the non-cooperative receiving of the NYFR is a wideband Nyquist subsampling process and the NZ index of the NYFR output is a key parameter to be estimated.Based on the Bessel function expansion of the NYFR non-uniform LO,the matching component function(MCF)is proposed to estimate the NZ index of the NYFR output complex signal.Compared with the existing method,the MCF can estimate the NZ indexes when signal modulation type is unknown and its estimation reliability is high.In addition,the computational load of the MCF is small and it can simplify the following signal processing of the NYFR output.The MCF is suitable for the mono-pulse(MP)signal,the linear frequency modulated(LFM)signal and the binary phase coded(BPC)signal.The relationships between the NZ index estimation performance of the MCF and the NYFR non-uniform LO parameters are analyzed quantitatively.The corresponding parameter selection criteria are given.Based on the MCF,a fast signal modulation identification method for the above three types of radar signals is proposed.According to the MCF,an extended discussion for the NZ index estimation of the multiple concurrent radar signals intercepted by the NYFR is analyzed.The effects of the NYFR and signal parameters on the corresponding NZ index estimation performances are analyzed.4)Based on the periodicity of the NYFR non-uniform LO,the LFM/BPC complex modulated radar signal intercepted by the NYFR is modeled as a periodic matrix after square processing.According to the proportional relation of the elements in different rows of the periodic matrix,the chirp rate SVD ratio(CSVR)spectrum is proposed to get the accurate chirp rate estimation result.In addition,the effect of the signal parameter on the estimation performance of the CSVR spectrum is analyzed theoretically.Combing with the NZ index estimation using the MCF and the demodulation process,the NYFR output signal is divided into several data segments with the same length.According to the properties of matching code and subspace pseudo-spectrum,the phase searching pseudo-spectrum is proposed,which can estimate the BPC modulation part with high accuracy.At last,the algorithm steps for the parameter estimation of the LFM/BPC signal intercepted by the NYFR are given and the proposed method can obtain a high accurate estimation result.