The combination of Kalman filtering technology and artificial neural networks technology has provided a new method for forecasting the changes of track geometry state. This article has established a forecasting model for the track geometry state of BP neural network based on Kalman filtering. The author analyzed some track detecting data of Chengdu-Kunming railway line of Kunming Railway Bureau, and made the state forecast based on this model. Simulation result showed that the application of BP neural network algorithm based on Kalman filtering in forecasting rail state has very high accuracy. The experiment showed that the application of this technology could forecast changes of the track geometry state in advance, and could make in-time adjustment to the track maintenance service work. Thus it is an effective measure for maintaining track geometry state in the best condition.

This content is only available via PDF.
You do not currently have access to this content.