Multi-sensor based Train Localization and Data Fusion in Autonomous Train Control System

It is well known that one of the key technologies in train control system is the localization of train. Since multiple sensors can be installed on the train to obtain real-time data, information fusion is a promising method that can be used to combine each sensor's unique information to calculate the stable and accurate localization results. A multi-sensor based localization system for the train autonomous control is proposed in this paper, which contains the Global Positioning System, Inertial Navigation System and velocity sensor. Covariance Intersection algorithm is proposed to integrate the output results of each sensor. Besides, considering the variety of the train running environment, adaptive Kalman filter is applied to reduce the impact of environmental noise. Finally, simulation results prove the proposed method in this paper improves the positioning accuracy compared with the traditional methods.