考虑高铁线路微观建模和问题知识的多列车驾驶策略实时决策方法OA

To further improve the intelligent decision-making level of the automatic train operation(ATO)system,this paper proposed a decision-making method for multi-train driving strategy that integrates microscopic modeling of high-speed railway lines and problem-specific knowledge.A microscopic model of the high-speed railway line was constructed by considering microscopic details,such as one-dimensional block sections and two-dimensional station details(throat targeting points,entrance/exit signals,turnouts,stopping targeting points,etc.).From the perspective of ensuring safe multi-train tracking,the optimization objective was designed as the total train arrival delay.By considering the multi-train tracking constraints in block sections and stations,a mathematical model for multi-train driving strategies decision-making was established based on the microscopic modeling of high-speed railway lines.The safe tracking process of multiple trains in block sections and stations was analyzed,and a multi-train driving strategy method was proposed based on tight tracking problem-specific knowledge.This method can effectively ensure accurate modeling of high-speed railway lines and efficient decoupling of strongly coupled spatiotemporal constraints among multiple trains.Furthermore,a virtual-coupling problem-specific knowledge method was designed to enable a more efficient real-time solution for multi-train driving strategies,satisfying the real-time requirements of ATO applications.A case study on the Beijing-Tianjin high-speed railway line was conducted under the typical temporary speed restriction scenario affecting nine trains.The results show that compared with traditional macroscopic railway line modeling methods,the proposed microscopic modeling method precisely generates multi-train driving strategies for block sections and stations.Compared with manual empirical methods,the proposed two problem-specific methods can both greatly reduce the total train delay,while the virtual-coupling problem-specific knowledge method dynamically adjusts multi-train tracking intervals.This effectively addresses the issues of the tight-tracking problem-specific method,such as redundant computation of driving strategies and frequent switching of train operating conditions,enabling real-time generation of multi-train driving strategies within 5.21 s to support decision-making for drivers.The research results can further optimize the precise stopping functionality of the automatic train operation system and enhance the intelligence decision-making level of train automatic driving.