Design of Robust Observers for Track-Roadbed State Monitoring

As the speed of railroad vehicles increases, the use of concrete tracks is increasing to ensure driving stability. However, continuous measurements of roadbed subsidence are required because concrete tracks have significantly decreased usability and high maintenance costs owing to roadbed subsidence. In addition, it is challenging to achieve real-time monitoring and continuous data collection using the existing data collection methods. In this study, an observer based on onboard measurements is proposed and designed for real-time monitoring. The proposed observer used the vehicle vibration signal and the suspension model of the railroad vehicle, and the error of the suspension model was compensated for using the dual Kalman filter. To verify the performance of the proposed observer, the virtual roadbed settlement is used as a reference value and the results are compared.