Research On Target DOA Estimation Of Automotive Millimeter Wave Radar

In recent years,with the development of economy and technology,the number of cars has increased sharply.As an important sensor,millimeter-wave radar can assist drivers in perceiving the surrounding links and provide decision-making basis for driving operations.However,the small number of array elements and small antenna aperture of the automotive millimeter wave radar result in poor angular resolution,and it is difficult to distinguish two targets with similar directions of arrival(DOA).Therefore,the application of millimeter waves has been Restricted.In this paper,the key parameters,estimation accuracy and resolution,of angle measurement are improved from two aspects: array layout and estimation algorithm.On the one hand,the sparse array design is carried out to efficiently use each element and increase the antenna aperture to achieve better directional pattern performance;on the other hand,the DOA algorithm research is carried out,and a simple and efficient DOA estimation algorithm suitable for the automotive millimeter wave radar platform is studied from the algorithm level to achieve High-precision and super-resolution DOA estimation.The main work of this paper is listed as followed.1.The basic principles and signal processing framework of automotive millimeter wave radar are sorted out,especially the principles of DOA estimation.Starting from the space-time relationship of the array signal,the characteristics and limitations of DOA estimation of real aperture and MIMO virtual aperture under the uniform linear array model are analyzed.2.Sparse array design.Firstly,two optimization algorithms,iterative Fourier algorithm and genetic algorithm,that can be used for array sparse design are introduced.Then,from the perspective of actual application scenarios,for the millimeter wave radar with 8 elements,the two algorithms are used respectively.A sparse array scheme that takes into account the side lobe level and the width of the main lobe is designed on 20 grids uniformly distributed at half wavelength.3.The high-precision DOA fast estimation algorithm is studied.The interpolation iterative Fourier algorithm is introduced into the DOA estimation field,and only fast Fourier Transform(FFT)and a small number of iterations can achieve fast and highprecision estimation of the target DOA,sharing the same order computational complexity with that of FFT.Theoretical analysis and experiments show that the algorithm is also effective for non-uniform linear arrays.4.A super-resolution DOA algorithm suitable for automotive millimeter wave radar is proposed.Utilizing the extremely high accuracy of the automotive millimeter wave radar,combining the continuous multiple measurement data of the moving target to synthesize a very large virtual array,which effectively realizes the angle super-resolution.Simulation experiments show that for 8-array millimeter-wave radars,even the traditional Digital Beam-Forming(DBF)algorithm can achieve angular super-resolution.5.Field experiments were carried out,and the actual millimeter-wave radar platform was used to collect data,and the measured data verification of the above-mentioned highprecision and super-resolution DOA estimation algorithm was completed.