作为国家战略基础设施之一，高速铁路在为我国工业化和城镇化建设发挥重要作用的同时，其在途运营的安全性保障也受到由恶劣多变的运行环境和高速多样的列车状态带来的诸多新挑战。其中，如何综合利用高铁运营大系统中的各类监控与检测数据，建立实时可靠异常状态预测和安全预警的有效机制，实现高效、可靠、安全的高速列车在途智能监测系统，就成为构建多层次、全方位的高速铁路安全运营保障系统的一个亟待解决的关键问题。本项目针对高速列车运行过程的主要特征，以建立集可靠性、实时性、安全性于一体的高速列车在途安全智能监测理论与方法为科学目标，综合运用深度学习、隐私保护、数据挖掘等多种技术，重点解决高速列车在途监测数据处理中的不确定异构信息整理与融合、多任务异常数据检测与分析、跨平台敏感内容保护与计算等关键问题。本项目的实施将为基于人工智能的高速列车在途安全监测奠定理论基础，提升我国在智能轨道交通领域的国际学术地位和影响

As part of the national strategic infrastructure, the high-speed rail plays a significant role in the industrialization and urbanization of the county. Meanwhile, the train safety operation faces the challenge of environment stresses and various train status. Thereinto, how to develop practical and real-time prediction approaches with big data from both railway systems and other sources is becoming a critical issue in establish the comprehensive safety monitoring system in high-speed train operation. This proposal aims at designing and developing novel theories and methods for intelligent safety monitoring in high-speed train operation. We focus on dealing with the key problem including uncertain heterogeneous data cleaning and fusion, outlier detection and analysis with multi-task learning, and privacy preserving data publishing in cross-platform. The success of this project will provide the fundamental results for safety monitoring in high-speed train operation, and elevate influence in intelligent railway systems.