A localization algorithm for railway vehicles based on sensor fusion between tachometers and inertial measurement units.

The availability of a reliable speed and travelled distance estimation is relevant for the efficiency and safety of automatic train protection and control systems. This paper investigates the main features of an innovative localization algorithm that integrates tachometers and inertial measurement units. Nowadays, the estimation is performed by an odometry algorithm that relies on wheel angular speed sensors. The objective is to increase the accuracy of the odometric estimation, especially in critical adhesion conditions, through sensor fusion techniques based on Kalman filter theory. The Italian company ECM S.p.A. has supported the project, providing a custom inertial measurement unit based on micro electro-mechanical system sensors for the on-track testing of the algorithm. The preliminary results show a significant improvement of the position and speed estimation performances compared to those obtained with SCMT (Italian acronym for ‘Sistema Controllo Marcia Treno’) algorithms, currently in use on the Italian railway network. A wide set of simulated test results, showing the improvement of the estimation process, is presented and discussed. An accurate train navigation that scarcely relies on information from the infrastructure will open a road map for the development of a more and more effective and efficient exploitation of the railway infrastructure. [ABSTRACT FROM PUBLISHER]