Estimation of Constant Speed Time for Railway Vehicles by Stochastic Gradient Descent Algorithm

Abstract

While the investments in rail transportation systems continue without slowing down, various optimization issues come to the fore in order for the systems to work more efficiently. One of the most important of these issues is the optimization of the vehicle speed profile. Improvement in vehicle speed profile increases efficiency in operating traffic. Vehicle speed profile varies depending on the electrical-characteristic features of the vehicle, the distance between the stations and the line geometry. The vehicle's speed profile consists of several parts, such as acceleration, constant speed travel and braking zones. The constant speed in the constant velocity zone refers to the max operating speed, which is recommended for operation in the restricted area and remains within the limits. This part is critical in creating the speed profile of the vehicle. In this study, the estimation of the value of the constant speed time in the speed profile of the vehicles used in the city metro systems was made by using the Stochastic Gradient Descent method, which is one of the machine learning methods, and compared with various well-known methods. Coefficient of determination (R²) values were calculated as 0.9955 and 0.9951, respectively, with random sampling hold out and cross validation methods.

Keywords

• constant
• railway
• speed
• stochastic
• vehicle

Supporting Institution

Istanbul Metropolitan Municipality, Rail System Department

Thanks

We would like to thank Istanbul Metropolitan Municipality, Rail System Department, for its support during the realization of this study.

References

1. J. Jong and S. Chang, "Algorithms for generating train speed profiles," Journal of the Eastern Asia Society for Transportation Studies, no.6, pp. 356-371, 2005.
2. Y. Chen and L.R. Rilett, "A train speed measurement and arrival time prediction system for highway-rail grade crossings," Transportation Research Record Journal of the Transportation Research Board, vol. 2608, no.1, pp. 96-104, 2017.
3. S. Hiraguri, "Evaluation of Train Control Method Using Prediction Control," Quarterly Report of Rtri, vol. 49, no. 3, pp. 163-167, 2008.
4. T. Ogawa, S., Manabe, G. Yoshikawa, Y. Imamura and M. Kageyama, "Method of Calculating Running Resistance by the Use of the Train Data Collection Device," Quarterly Report of RTRI, no.1 58, pp. 21-27, 2017.
5. S. Hensel and M. Marinov, "Time Signal Based Warping Algorithms for Low Speed Velocity Estimation of Rail Vehicles," Annual Journal of Electronics, no. 8, pp. 177 - 180, 2014.
6. K. N. Cosgriff, E. G. Berggren, A. M. Kaynia, N. N. Dam and N. Mortensen, "A new method for estimation of critical speed for railway tracks on soft ground," International Journal of Rail Transportation, vol. 6, no. 4, pp. 203-217, 2018.
7. G., Xu, F., Li, J., Long, D., Han, "Train movement simulation by element increment method", Journal of advanced transportation, no. 50, pp. 2060–2076, 2017.
8. S., Aradi, T., Becsi, P., Gaspar, "Estimation of running resistance of electric trains based on on-board telematics system", International Journal of Heavy Vehicle Systems, no. 22, pp. 277-291, 2015.

9. T., Kunimatsu, T., Terasawa, Y., Takeuchi, "Evaluation of Train Operation with Prediction Control by Simulation", in International Conference on Railway Operations Modelling and Analysis, 2019, pp.589-606.
10. E. Naye, "Real-time arrival prediction models for light rail train systems," Royal Instistute of Technology, Department of Engineering, Master’s Thesis, 2014.
11. Z. Lv, J. Xu, K. Zheng, H. Yin, P. Zhao and X. Zhou, “LC-RNN: A Deep Learning Model for Traffic Speed Prediction,” in Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence (IJCAI-18), pp. 3470-3476, 2018.
12. A. Dhamaniya and S. Chandra, “Speed Prediction Models for Urban Arterials under Mixed Traffic Conditions," Procedia - Social and Behavioral Sciences, no. 104, pp. 342 – 351, 2013.
13. M. Gmira, M. Gendreau, A. Lodi and J., Potvin, “Travel Speed Prediction Based on Learning Methods For Home Delivery,” Interuniversity Research Center On Business Networks, logistics and transport, pp. 1-34, 2018.
14. M. Bysveen, “Vehicle speed prediction models for consideration of energy demand within road design,” Norwegian University of Science and Technology, Civil and Environmental Engineering, Master’s Thesis, 2017.
15. B. Mirbaha, M. Saffarzadeh, S. A. Beheshty, M. Aniran, M. Yazdani and B. Shirini, "Predicting Average Vehicle Speed in Two Lane Highways Considering Weather Condition and Traffic Characteristics," IOP Conference Series: Materials Science and Engineering, pp. 1-7, 2017.
16. M. Gmira, M. Gendreau, A. Lodi and J. Potvin, "Travel speed prediction using machine learning techniques," ITS World Congress, pp. 1-10, 2017.
17. O. Cats, "Real-Time Predictions for Light Rail Train Systems," 17th International IEEE Conference on Intelligent Transportation Systems (ITSC), pp. 1-10, 2014.
18. G. Kouroussis, D. P. Connolly, M. Forde and O. Verlinden, "Train speed calculation using ground vibrations," Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, vol. 229, no. 5, pp. 466-483, 2015.
19. R. Kaleem, S. Pai and K. Pingali, “Stochastic Gradient Descent on GPUs,” ACM International Conference Proceeding Series, pp. 81-89, 2015.
20. I. Chakroun, T. Haber and T. Ashby, “SW-SGD: The Sliding Window Stochastic Gradient Descent Algorithm,” Procedia Computer Science, no. 108, pp. 2318-2322, 2017.
21. J. Keuper (Fehr) and F. J. Pfreundt, “Asynchronous parallel stochastic gradient descent: a numeric core for scalable distributed machine learning algorithms,” MLHPC '15: Proceedings of the Workshop on Machine Learning in High-Performance Computing Environments, pp. 1-11, 2015.