Research On Vehicle-Mounted Millimeter Wave FMCW MIMO Radar Target Tracking Technology

With the rapid development of information intelligence technology,intelligent driving technology has become a popular research direction,driving the development of advanced driver assistance systems(ADAS).Millimeter wave radar has the advantages of working in harsh environmental conditions and maintaining good robustness,as well as high resolution and small size,making it one of the core components of ADAS.The target tracking technology of in-vehicle millimeter wave radar is an extremely important function in ADAS,which can obtain and analyze the road target information in real time,and can control the car accordingly in case of emergency to improve road safety and reduce the traffic accident rate.This paper investigates the key issues of millimeter wave radar target tracking based on a 77 GHz vehicle-mounted millimeter wave frequency modulated continuous wave(FMCW)multiple input multiple output(MIMO)operating regime radar,mainly.(1)A detection method based on millimeter wave radar for moving targets is proposed for the problem of Doppler blurring of targets.A velocity expansion method for radar velocimetry by improving the Chinese residual theorem(CRT),which improves the robustness and velocimetric accuracy of the CRT in the velocity ambiguity case.The Doppler phase shift is corrected by the phase compensation method on the phase shift,and then the target angle information is accurately estimated by the FFT beamforming method with Kaiser window,which can effectively solve the angle ambiguity problem generated by the time division multiplexing(TDM)multiple-input multiple-output(MIMO)regime radar when there is relative motion with the target.Simulation and experimental results verify the effectiveness of the proposed algorithm.(2)In response to the problem of uncertainty in the source of measurement data when tracking multiple targets in a simple environment with low traffic flow and a single road condition,the research of target tracking technology based on data correlation is carried out.Firstly,three motion models for target tracking and the associated wave gate algorithm are introduced.Then the extended Kalman filter(EKF),particle filter(PF),and traceless Kalman filter(EKF)methods are studied and simulated and analyzed.Finally,the Joint Probabilistic Data Association(JPDA)tracking algorithm is used to correlate the radar observed data with the target’s track,thus solving the problem of mismatch between the actual target data of the echo and the track due to the uncertainty of the source of the measurement data when tracking multiple targets,and the effectiveness of the algorithm is verified by simulation.(3)The research on random finite set(RFS)based target tracking technology is carried out to address the problem of unsatisfactory multi-target tracking by traditional tracking techniques in a complex environment with high traffic flow and chaotic road conditions when crossings are gathered in clutter.Firstly,the definition of RFS and the filtering model are described,and the mathematical derivation of the RFS filtering model formula is presented.Then,the target state and measurement are constructed based on RFS using nonlinear system extended Kalman filter Gaussian mixed probability hypothesis density(EKF-GMPHD)to represent the process of target’s newborn and disappearance,and to update the prediction,thus solving the problem of tracking degradation in complex environment due to target gathering cross in clutter.The simulation results verify the effectiveness of the algorithm.