Research on the Method of Extracting Safety Constraints from On-Board ATP Operation Data Based on CNN

Each time a train completes a running route on the main line, the operation data generated by the on-board ATP system will form a data plane, and the features presented in this data plane are the safety constraints of the on-board ATP operation data. The paper analyzes some on-board ATP data variables and uses the change characteristics of data variables as the extraction conditions of data sections. By extending the dimensions of data variables, the preprocessing of operation data is achieved, and a hierarchical dynamic data model of on-board ATP operation data is established. Using Convolutional Neural Network (CNN) to perform convolution operations on the data of hierarchical dynamic data models to extract data features, meaningful features can be accurately captured from the data, and safety constraints for on-board ATP operation data can be provided.