Monitoring the parameters of high mobility trains are required for the safe navigation and reliable communication applications of the high speed vehicle. The localization of the train accurately is a challenging task. Hence an effective model and an efficient algorithm must be used. Some of the traditional methods of locating the position of the high speed train include the use of GPS. However, GPS cannot always give precise location as other external factors like climate and mobile environment also plays a major role. So, there is a need for localization of the train by itself. Thus, extra sensors are needed which can increase the precision of localization. Localization of high speed train not only falls into the domain of safety enhancement but also falls into the domain of designing smart trains which are energy efficient and intelligent. This information is also useful for establishing a real time wireless communication link in high mobility system. In this paper, we proposed an algorithm which involves estimating two angles, pitch angle and roll angle in addition to the velocity of the train from the noisy sensors like gyro-meter and accelerometer. This problem falls under the conventional non-linear estimation problem. Choosing the correct estimation algorithm is as important as correctly modeling the train dynamics. Existing literature made use of the extended Kalman filter (EKF) to estimate the state variables accurately for different physical models. In this paper, we propose to use the unscented Kalman filter (UKF). Theoretically, UKF is a better estimator compared to EKF because it captures higher order statistics of the system. The simulation results show that the proposed UKF based tracking algorithm is a reliable estimator by comparing the estimation error of UKF algorithm and EKF algorithm for estimating the pitch angle, roll angle and velocity of the train.
P. Sudheesh and Dr. Jayakumar M., “Tracking of Nonlinear Variations of the Parameters of High Mobility Systems”, International Journal of Pure and Applied Mathematics, Open Access, vol. 118, no. 7, pp. 221-226, 2018.