A Multi-Sensor Interacted Vehicle-Tracking Algorithm with Time-Varying Observation Error

Vehicle tracking in the field of intelligent transportation has received extensive attention recent years. Multi-sensor-based vehicle system is widely used some critical environments. However, actual scenes, observation error each sensor often different and time varying because environmental change channel difference. Therefore, this paper, we propose a multi-sensor interacted vehicle-tracking algorithm with time-varying (MI-TVOE). The establishes jointed model for to indicate variation noise. Then, develop which can predict statistical information fuse result provide global estimation. Simulation results show that proposed MI-TVOE significantly improve performance compared single-sensor-based method, traditional unscented Kalman filter (UKF), apdative UKF method (AUKF) multi-error fused (MEF-UKF), will be well applied complex scenes reduce computational complexity error. experiments paper also prove superiority