Research On Stable Multi-target Tracking Algorithm Of Vehicle-borne Millimeter Wave Radar

In recent years,automotive assisted driving technology has been widely used,especially in the prevention of traffic accidents.In the automobile-assisted driving system,the millimeter-wave radar is an important part.It can play a role in bad weather and complicated road conditions,and has a low cost and a wide application prospect.An important function of the car-assisted driving system is the tracking of maneuvering targets.Currently,there are many algorithms to solve this problem.The traditional data association algorithm is ideal for tracking under complex road conditions,but excellent tracking accuracy is required for sensors and other components.High,at the same time,when the number of targets increases,the difficulty of calculation increases significantly.The modern multi-target tracking theory based on stochastic finite set deduces through strict probability theory and bypasses the data association link,which significantly reduces the computational cost and is also very suitable for practical applications.The main work of this article is as follows:(1)The working principle of several common vehicle-mounted millimeter-wave radars is studied and their advantages and disadvantages are compared.Then the dualfrequency continuous-wave radar was modeled and the multi-target detection method of dual-frequency continuous wave radar was discussed.Simulations were performed under typical operating conditions.Ground clutter interference on complex roads is discussed during car driving and simulation analysis is carried out.(2)The traditional Kalman filter algorithm is analyzed.Because the tracking target of the vehicle radar contains a maneuvering target,the maneuvering target tracking method is studied and simulated.(3)Researching the traditional multi-target tracking method based on data association,analyzing the advantages and disadvantages of the joint probability data association algorithm,performing multi-objective simulation on it,and combining this algorithm with the interactive multi-model algorithm in the maneuvering target Tracking has achieved ideal results.(4)The feasibility of multi-target tracking method based on stochastic finite set theory in practical application is given,and the simulation of Gaussian mixture probability hypothesis density algorithm is implemented.The improvement of the algorithm in the presence of strong interference in the presence of strong interference is proposed.Measure,by combining the nuclear density algorithm to improve the tracking effect,gives the steps of the algorithm and simulation.