original research article
Train Localization using Unscented Kalman Filter–Based Sensor Fusion
Ismail Faruqi, Muhammad Brahma Waluya, Yul Yunazwin Nazaruddin, Tua Agustinus Tamba
This paper presents an application of sensor fusion methods based on Unscented Kalman filter (UKF) technique for solving train localization problem in rail systems. The paper first reports the development of a laboratory-scale rail system simulator which is equipped with various onboard and wayside sensors that are used to detect and locate the train vehicle movements in the rail track. Due to the low precision measurement data obtained by each individual sensor, a sensor fusion method based on the UKF technique is implemented to fuse the measurement data from several sensors. Experimental results which demonstrate the effectiveness of the proposed UKF-based sensor fusion method for solving the train localization problem is also reported.