Study On Signal Processing For Automotive Millimeter Wave Radar

As the development of social economy and the improvement of living level,the number of cars per capita increases rapidly,the regulation and safety of road traffic become much more important,which leads to the tremendous development of autonomous driving technology in recent years.At present,millimeter-wave radar has become one of the most important sensors in the field of autonomous driving due to its small size,reasonable cost,and stable performance under severe conditions.This thesis mainly studies the signal processing of vehicle-mounted millimeter wave radar.And a set of relatively complete automotive radar target detection algorithm has been designed.We first analyzes the problems that FMCW radar needs to solve when estimating target distance,speed and angle,and then focus on the Range-Doppler-Angle 3D-FFT target detection algorithm with fast chirp signal as the transmission waveform for the optimal detection of vehicle-mounted millimeter wave radar Program.And it is verified by Matlab simulation that the 3D-FFT detection algorithm can effectively avoid the distance-velocity coupling problem of the traditional large-time-width sawtooth radar waveform and the multi-target speed matching problem of the triangular waveform.In addition,the 3D-FFT algorithm can significantly improve the spatial resolution of the target positioning compared with the traditional single-pulse and beamforming methods.And compared with the super-resolution algorithm with higher angular resolution,the 3D-FFT detection algorithm has a smaller amount of calculation,can obtain the strength information of the target echo and as well as without the need for prior information about the number of targets.Due to the alternative transmission characteristic of the traditional Time Division Multiplexing-Multiple Input Multiple Output(TDM-MIMO)radar,the sampling rate of the velocity dimension of the echo signal is reduced,which results in the detection range of the target speed is also reduced.Therefore,a new TDM-MIMO radar antenna layout scheme is designed regarding to 2 transmitting and 4 receiving antennas.It is proved by Matlab simulation that the 3D-FFT detection algorithm based on the antenna layout scheme can double the maximum detection speed compared with the existing design.Finally,we conduct real target detection experiment by using the IWR1443 radar development board and DCA1000 real-time data capture adapter from Texas Instruments so as validate the developed 3D FFT algorithm.The experiment is designed for different scenarios,e.g.a pedestrian as a slow target while a motor / car as a fast target.And the results show that our algorithm works well: the range,angle as well as the speed of targets can be well estimated.