Knowledge Based MIMO Radar Waveform Design And Wideband Radar Detection

With the rapid development of information and digitization,facing the complex and changeable external scene,how to have more flexible scanning mode and working mode,how to obtain more target information and higher measurement accuracy,and how to effectively deal with all kinds of active and passive jamming are the main development direction of radar.Nowadays,multiple input multiple output(MIMO)system and large bandwidth become two choices for radar performance improvement: MIMO radar has multiple transmit and receive antennas,and each antenna unit transmits different signals to ensure higher degree of freedom and measurement accuracy;while wideband radar has the characteristics of large bandwidth and high range resolution.For MIMO radar,transmit waveform and transmit beampattern are important factors affecting its performance,while for wideband radar,moving target detection in complex environment is one of the factors restricting its performance.The transmit beampattern is physically the distribution of radar power in space,which will directly affect the intensity of radar echo;under the premise that the transmit antenna array is fixed,it is mathematically determined by the covariance matrix of the transmit waveform.In order to flexibly control the transmit beampattern and realize the efficient use of radar system resources,it is necessary to optimize the design of the transmit waveform covariance matrix,or the transmit waveform;in addition,the radar echo can be regarded as the return after the transmit waveform interacts with the target and the environment.The separation of strong and weak targets also requires a reasonable design of the waveform.Due to the large emission bandwidth of wideband radar,the traditional point target model and Gaussian clutter model are no longer applicable.It is necessary to model the wideband radar target and clutter to design a detector suitable for the wideband radar.This dissertation is based on this realistic demand,focusing on MIMO radar low sidelobe transmit beampattern design,cognitive waveform design,spatial-temporal optimization;wideband radar target detection in a heterogeneous clutter environment,and wideband MIMO target detection are studied.The specific work of this dissertation can be summarized into the following four parts:1.For the transmit waveform beampattern design of MIMO radar,three different methods are proposed.The first two methods are based on the optimization of covariance matrix and only focus on the spatial beampattern approximation.The latter method directly optimizes the transmit waveform to achieve the beampattern matching in the spatial and the autocorrelation in the temporal.The first method is a low-sidelobe transmit beampattern design method,which introduces two parameters,the upper and lower bound of the mainlobe,and the lower bound of the mainlobe is fixed,then a semi-positive quadratic programming convex optimization mathematical model is designed.By controlling the upper bound of the main lobe,the main lobe fluctuations and low side lobes of the launch pattern are adjusted simultaneously.The second method is to transform the beampattern matching problem into a matrix rank minimization problem,then the matrix rank minimization as the optimization objective is given.Using matrix transformation,the multi-dimensional signal model is converted into a one-dimensional signal vector,so that the rank of the covariance matrix is1.Take matrix reduction as the objective function,the mainlobe fluctuation,sidelobe,semipositive covariance matrix are constraint,and finally use the log-det heuristic algorithm to solve it.The third method is to directly optimize the transmit waveform.Firstly,an improved sequential convex programming algorithm is used to optimize the intermediate signal to achieve low sidelobe transmit beambattern.Secondly,the gradient method is used to optimize the diagonal phase matrix to achieve low autocorrelation sidelobes of the spatial synthesized signal.Finally,the intermediate waveform and phase matrix are combined to obtain a constant mode transmit waveform that is close to the desired beampattern.2.Based on the concept of cognitive radar,MIMO radar receiver feedback is used to adjust the transmitter to achieve online waveform design and generation,so as to achieve the purpose of clutter suppression and target detection.Firstly,the multi pulse multi-target echo model of MIMO radar is established,the heterogeneous clutter model related to the waveform is designed,and the sampling covariance matrix is obtained.According to the NP criterion,the detection probability is positively correlated with the generalized likelihood ratio when the false alarm probability is fixed.Therefore,using the prior information of clutter statistics,the objective function of generalized likelihood ratio is established under the Bayesian framework related to the transmit waveform.At the same time,in order to make the radar work at the maximum efficiency,it is necessary to set the constant modulus of the transmit waveform,so that the optimization problem is non convex.Then two methods are proposed to solve the original problem.The first method uses ADMM algorithm,which has high complexity and time-consuming,but has good clutter suppression effect.The second method uses phase matrix relaxation to optimize the original problem and BFGS algorithm to solve the relaxation problem.The advantage of this method is that the complexity of the algorithm is low,and the disadvantage is that the clutter suppression performance is slightly worse than the former method.The simulation results show that the waveform designed can improve the performance of clutter suppression and target detection.3.The design of knowledge aided Bayesian detector for wideband radar in 2-D compound Gaussian correlated clutter in pulse slow time dimension and range frequency domain dimension is proposed.Firstly,the accurate model of multi coherent pulse echo of wideband radar target considering the range cell migration is introduced,and the linear target echo model related to the target amplitude is derived by discrete Fourier transform.Then,the twodimensional wideband radar clutter model and its prior information are modeled.Then,the classical model of detection problem is established.The density probability function of the observation data is obtained by Bayesian theorem integration,and the numerical analytic expression of the second confluence hypergeometric function is obtained by Laplace approximation.Based on this,the GLRT detector,Rao detector and Wald detector for wideband radar target are designed.Monte Carlo simulation results show that the Rao detector has constant false alarm rate(CFAR)property on the 2-D clutter covariance matrix.If the correlation of clutter in range frequency domain is small,Wald detector also has CFAR characteristics.In the case of small samples,the designed detector can approach the optimal detector,and for sparse target detection,the performance of the three detectors is very stable.4.The method of knowledge aided target detection for wideband MIMO radar based on secondary data is studied.Firstly,the linear echo model of wideband MIMO radar considering the cross antenna aperture and range cell migration is derived,and the heterogeneous clutter model which obeys the compound Gaussian model is introduced.In order to model the clutter flexibly,the Wishart distribution is used as the prior of the clutter covariance matrix,and the covariance matrix of the clutter training data obeys the inverse Wishart distribution with the clutter covariance matrix as the center matrix.When the clutter power obeys the inverse gamma distribution,the joint detection problem of observation data and training data is established.However,the accompanying problem is that the optimal generalized likelihood ratio detector cannot be calculated directly,so the IG-MAP-GLRT detector is designed by using inequality approximation.Simulation results show the CFAR characteristics and effectiveness of the designed detector.