Research on Risk Accurate Prediction Method of Train Platoon Operation Based on KPCA-EBWO-SVM

In response to the problems of low utilization of monitoring data for platoon train operation, low accuracy in risk identification and prediction, as well as poor initiative in emergency response, a support vector machine hybrid prediction classification method based on KPCA and EBWO is proposed in this paper. Firstly, the multi-factorial nature of the risk encountered during operation in platoon is analyzed, and system risk levels are defined based on the operational status of platoon. Secondly, the KPCA method is used to perform data dimensionality reduction and feature extraction on operational data, quasi reverse learning and whirlwind foraging strategies are utilized to continuously update SVM model parameters, so as to achieve efficient processing of complex data and accurate prediction and classification of risk types. Finally, based on actual train parameters and simulated operation data, the effectiveness of the proposed method in improving the utilization rate of operation monitoring data, risk identification and prediction classification accuracy is verified. In turn, it provides a guarantee for the intelligent, reliable and safe operation in flexible platoon of rail transit.