An Improved Multi-target Tracking Algorithm for Automotive Radar
Fan Huang, Jianjiang Zhou, Xiaotong Zhao · Journal of Physics Conference Series · 2021
Abstract An improved multi-target tracking algorithm for automotive radar is proposed in this paper. In the application of automotive radar, in addition to the motion generated by the target maneuver, the maneuver of the radar-equipped vehicle will also cause the target to move relative to the radar. A single motion model for the target cannot accurately describe the motion state of the target relative to the radar. To solve the above problems, the nonlinear state equations of various motion models and their process noise covariance matrix are derived, and the Interacting Multiple Model (IMM) algorithm, the Joint Probabilistic Data Association (JPDA) algorithm and the Unscented Kalman Filter with Doppler measurement (DUKF) are combined. At the same time, the correlation coefficient of range measurement error and velocity measurement error of Frequency Modulated Continuous Wave (FMCW) radar is derived, and the relationship between it and FMCW radar parameters is studied. By simulation, its influence on target tracking is assessed, final results indicate the reference value for the application of automotive radar in intelligent driving.