Research on Multi-Sensor Information Fusion for Positioning of Suspended Maglev Trains based on DOO Weighted UKF

The accuracy deficiencies, complex computation and single sensor of the filter algorithm used in the positioning system of suspended magnetic levitation train (SMT) seriously affect the precise positioning of maneuver targets. Leveraging the system's observability, this paper presents an unscented Kalman filtering (UKF)-based multi-sensor information fusion method for positioning suspended magnetic levitation train. The Degree of Observability (DOO) is used to evaluate and analyze the information quality and working performance, and the weights of the observations of the levitated magnetic train are assigned according to the DOO value, and the new observations obtained are weighted fused using the UKF. Experimental results demonstratethat the use of UKF based on observability and multi-sensor information fusion offers considerable accuracy and simplicity advantages over single-sensor and random weighting. The results establish a theoretical basis for addressing the SMT positioning problem and provide crucial theoretical support for the advancement of maglev positioning technology.