Research On Vehicle Front Target Detection And Tracking Algorithm Based On Millimeter Wave Radar

With the development of millimeter wave radar technology and production technology,the size of radar is getting smaller and smaller,and the measurement accuracy is getting higher and higher,making automotive radar one of the irreplaceable sensors in advanced assisted driving technology and autonomous driving perception technology.It can be used to improve the safety and driving comfort of vehicles and avoid the loss of casualties caused by vehicle collisions.The millimeter-wave radar uses radiated electromagnetic energy to measure targets in the sensor's field of view,and has a long-range target detection capability.Compared with other automotive sensors,automotive radar can provide the target's unique speed(Doppler)measurement.It still has better robustness in severe weather and strong light environments,and the cost is relatively low in general.Based on the 77 GHz TDM-MIMO system millimeter-wave radar,the original echo signal is time-frequency transformed to extract the information of the detection target,on this basis,the tracking algorithms of point target and extended target are studied.The specific research work of this paper is as follows:(1)Radar mainly uses multiple input multiple output technology(MIMO)to increase the number of virtual antennas to improve angular resolution.For the theoretical part of the algorithm,this paper analyzes the theory of linear continuous frequency modulation sawtooth wave ranging and speed measurement under the TDM-MIMO system,and derives the signal model and virtual array of the TDM-MIMO radar.During relative movement,the amount of phase change caused by the Doppler frequency of the moving target during the switching time of different transmitting antennas will be coupled to each receiving antenna,resulting in a defocus effect in the frequency spectrum in the angular dimension.(2)The detection algorithm process is designed to solve two types of errors that often occur in radar detection.The first is the amplitude and phase errors caused by other hardware factors such as production processes or mutual coupling between antennas.In view of this problem,an active calibration scheme is designed to calibrate the amplitude and phase consistency of the signal to improve the angle measurement reliability.The second is that there is a coupling between the Doppler velocity and the measurement angle.In view of this problem,a phase compensation method is proposed by analyzing the cause of the phase error of the moving target,and a corresponding solution to the speed fuzzy case is designed.This method does not need to estimate the target speed and does not require additional hardware overhead.Finally,the feasibility of the detection algorithm is proved by actual measurement data,and it has low time complexity.(3)In order to better track point targets in a complex multi-target environment,the extended Kalman filter algorithm(EKF)is used to improve the density-based clustering algorithm(DBSCAN),and verified through simulation and measured experiments.The results show that the new algorithm can maintain a stable and low level each time when performing incremental clustering,and perform adaptive clustering without increasing the time complexity,which can solve the unevenness of automobile radar data density.Case.It can be seen that the new algorithm realizes both incremental and adaptive DBSCAN clustering,while ensuring the efficiency and accuracy of clustering.(4)To solve the currently difficult extended target tracking problem,instinctively add Doppler speed to the current state-of-the-art random finite set theory tracking method,and use Poisson multi-Bernoulli Mixture(PMBM)algorithm to predict the target Updated,designed a rectangular model,corrected the abnormal Doppler velocity,and used the Gibbs sampling method for data correlation,which can greatly reduce the calculation amount and substantially improve the tracking efficiency and tracking accuracy.