Research On Long-time Coherent Integration Method For Maneuvering Target Based On Multi-linear Kernel Of Xidian University

In the modern battlefield,radar targets have the characters such as high speed,high maneuverability,and low detectability.It is known that prolonging the integration time improves radar detection performance for weak targets.However,in a long integration time,the Doppler spectrum of a high-order motion target is broadened,and a high-order range migration occurs.The range migration and Doppler migration defocus the integration result and worsen estimation and detection performance.This dissertation considers both the theoretical and practical problems.First,we look into the signal model of a maneuvering target,the polynomial phase signal,and then propose a parameter estimation method for high-order polynomial phase signal.Second,we study on the integration algorithm for maneuvering target based on multi-linear kernel and propose three integration methods for maneuvering targets.The main content is as follows.1.In the first part,an estimation method is proposed based on iteratively reducing-order technique for a high-order polynomial phase signal.The proposed method reduces the SNR loss of traditional multi-linear kernel based algorithms.In the proposed method,we define a reducing-order operator.By combining the newly defined operator and phase differentiation operator,different kinds of multi-linear kernels are developed depending on the order of original polynomial phase signal.Basing on these multi-linear kernels,corresponding parameter estimation methods are proposed.The nonlinearity analysis,computational complexity analysis,and theoretical mean square error analysis are provided.Numerical examples are given to validate the effectiveness of the proposed method.The results of the experiments show that the proposed method can be performed efficiently by using fast Fourier transform.Moreover,compared with traditional multi-linear kernel based algorithms,the proposed method has lower order of nonlinearities and better anti-noised performance.2.In the part two,we consider the high computational complexity problem of the searching based integration algorithms.Combining the new defined reducing-order operator in Section III and the keystone transform,a fast integration algorithm for third-order maneuvering target is proposed.We define a kind of new multiliear kernel,which can be used to remove the third-order range migration and Doppler migration.Basing on the proposed multilinear kernel and two-dimensional fast Fourier transform,a new parameter estimation method for third-order maneuvering target is developed.We optimize the proposed method,and provide the cross-term analysis,computational analysis,SNR loss analysis,theoretical mean square error analysis and real data processing results.The experiments for single and multiple targets are also carried out.The results of experiments and analysis show that the proposed method can effectively eliminates the range migration and Doppler migration caused by target's third-order motion.The proposed method is applicable for high speed target and multiple targets scenes.The computational complexity is at the same level with the ACCF(adjacent cross correlation function)based algorithm.Moreover,because of its lower order of nonlinearities,the proposed method has lower SNR threshold and MSE compared with the ACCF based algorithm.3.In the third part,we consider the low anti-noise performance problems of the traditional multi-linear kernel based algorithms,and propose a new integration method with high anti-noise performance for third-order maneuvering target.In this part,we develop a new multi-linear kernel which includes multiple integration variables.Using the new defined kernel,we further proposed an integration method for a target with third-order motion,which has a good anti-noise performance.This method combines phase differentiation operator and keystone transform,and can eliminate the range migration and Doppler migration caused by target's first and second motion.After compensating the phase error caused by target's third-order motion,the target energy can be integrated by summing the signal data in different range frequency cells and performing two-dimensional FFT.Computational complexity analysis,cross-term analysis,simulations,and real-data processing results are provided.The analysis and experiment results show that the proposed method can effectively compensate the third-order range migration and integrate target's energy.Since the multiple integration variables are employed in the multi-linear kernel,the proposed method has a good anti-noise performance and cross-term suppression ability.The proposed method is applicable for low SNR environment and multi-target scene.Compared with the method proposed in Section III,the computational complexity of the proposed method grows,however is still lower than traditional searching based algorithms.4.In the part four,in order to achieve a balance between the anti-noise performance and computational complexity,we propose a parametric designing method for multi-linear kernels,which can effectively transform a polynomial phase signal to a complex sinusoid.Then,the energy of a polynomial phase signal can be integrated by performing keystone transform and multi-dimensional FFT.The principles of the parameter selection and designing method are provided.By using the proposed method,we develop two integration algorithms for respectively second and third order maneuvering targets.The computational complexity,and cross terms are analyzed.Simulations are provided for single and multi-target scenes.The results show that the proposed methods can effectively integrate the energy of maneuvering target,have good cross-term suppression abilities,and are applicable for low SNR environment and multi-target scenes.