Research On Multi-target Detection Algorithm For Millimeter Wave FMCW Radar

In recent years,the multi-target detection of millimeter wave FMCW Radar in complex environment has gradually become a research hotspot in many industries.Aiming at the problems of spectrum overlap,false target,difficult to distinguish dense frequency and limited frequency estimation accuracy of millimeter wave FMCW Radar in multi-target scene,this thesis discusses the multi-target matching algorithm in signal processing module The main contents are as follows:In recent years,the multi-objective detection of millimeter wave FMCW radar in complex environment has become a hot topic in many industries.Aiming at the problems of spectrum overlap,false targets and dense frequency resolution and limited frequency estimation accuracy in multi-objective scene,this thesis presents a multi-objective matching algorithm The preprocessing algorithm and target frequency estimation algorithm of target frequency estimation are studied.The specific work contents are as follows:(1)The research of multi-objective matching algorithm.Aiming at the problem of spectrum overlap of difference frequency signal caused by range velocity coupling and false target caused by target spectrum matching error in multi-target scene,this thesis analyzes the advantages and disadvantages of sawtooth wave and triangular wave frequency modulation in range velocity decoupling,compares the performance of triangular wave and trapezoidal wave frequency modulation in eliminating false target,and finally combines the idea of variable period frequency modulation,An improved multi-target pairing algorithm based on variable period triangular wave and constant frequency continuous wave is proposed.Firstly,the target spectrum pairs are filtered for the first time in the constant frequency band to eliminate most false targets,and then the false targets are removed again according to the characteristics that the real target parameters do not change with the modulation period parameters,Simulation results verify the stability and low complexity of the algorithm.(2)The research on the preprocessing algorithm of multi-target signal frequency estimation.Aiming at the problem of serious spectrum leakage in dense spectrum,the characteristics of three kinds of common window functions are analyzed and discussed.The experimental results show that Blackman window has the most obvious side lobe attenuation in dense spectrum,and the amplitude precision is the highest;Aiming at the problem of accurate target peak extraction in the spectrum of FFT processed difference frequency signal,the performance of mean class,ordered class and derived CFAR detection algorithm is compared and analyzed in different environments.The experimental results show that TM-CFAR can improve the problem of target peak masking and false alarm and missing alarm at clutter edge in frequency domain detection,and is less affected by the number of interference targets,It can accurately extract the target frequency peak.(3)The research on frequency estimation algorithm of multi-target signal.Aiming at the problem that it is difficult to distinguish the close frequencies in the dense spectrum and that the low accuracy of frequency estimation results in the large error of target range velocity measurement,a Zoom-MRife algorithm is proposed.Firstly,the Zoom FFT band-pass filter is used to refine the spectrum near the target peak to improve the local frequency resolution,and then the M-Rife algorithm is used to estimate the frequency deviation of the refined spectrum,The experimental results show that the algorithm has low root mean square error and can distinguish dense frequencies.