受限条件下列车自适应容错定位方法研究

GNSS-based train localization can be applied in the modern train control system to improve the performance of train localization. The trackside localization equipment can be thus reduced to save the implementation and maintenance cost. The satellite navigation-based train control has been considered as an important trend for the modernization development of railway transportation and the implementation of new generation train control systems. As train normally runs with long range, the position in space and the operation environments vary frequently with time. However satellite signals are prone to be obstructed or interfered during much of the scenarios, like tunnel, rail cutting, canopy area etc. The train positioning performance would be degraded when runs in such area. It is hard to achieve seamless train localization depending on solely GNSS positioning method. This project aims to fuse the multi-GNSS, INS, odometer and digital track map and construct a multi-sensor train positioning system so as to fulfil the rail safety requirements. The project aims to design the adaptive switched integration system scheme according to the different application scenarios. In order to achieve the fault tolerable rail positioning in the complex environments, especially the GNSS-denied area, an adaptive measurement statistics model would be applied in the integration system. Meanwhile, a new information space based data fusion algorithm would also be applied in the multi-sensor positioning system to help achieve a higher positioning solution. The research outcomes of this project will provide an effective solution to the corresponding technical problems in train operation field, promote the implementation of the GNSS in railway train control and other safety-related areas, and improve the intelligence and independence in the development of railway transportation systems.

基于卫星的列车运行控制系统可以减少轨旁的定位设备从而实现连续的实时列车定位，是铁路系统智能化以及新型列控系统用于铁路系统建设发展的重要趋势。列车在运行过程中空间位置及所处的环境随时间不断变化，运行环境复杂多变，而卫星定位在隧道、路堑、雨棚等线路地区易受到遮挡或干扰，依靠单一卫星定位信息源很难实现连续无缝的列车定位。本课题从列控系统对卫星定位单元的安全需求出发，基于多星座多频卫星导航系统 (GNSS)、惯性导航系统 (INS)、里程计以及轨道地图数据库构建多传感器列车定位系统，根据场景设计可切换组合系统方案，同时综合利用自适应量测统计模型和基于信息空间的最优信息融合方法，实现复杂场景下，特别是GNSS受限场景下的列车可靠容错定位，力求在列车自主定位的精确性、可靠性、连续性、可用性、完整性等性能上实现质的飞跃，从而有力支持众多基于位置的轨道交通列车运行控制的上层应用。

基于卫星导航的列车定位技术是实现自主、高效的新型列控系统发展的关键技术之一。然而，将卫星定位引入具有安全苛求特征的列车控制系统，面临卫星快速准确定位、单一卫星信息源受限、卫星信号故障等问题带来的挑战。本项目从基于卫星导航的列车运行控制系统实际需求入手，从实时高精度定位、连续无缝定位、自适应鲁棒定位三个方面，深入研究了列车在受限环境下连续无缝定位所需要的相关方法。项目研究工作取得的主要成果可概括为三个方面：1）提出了使用了列车运行的多星座多频列车快速定位方法，分别基于BDS/GPS双星座联合、实时非差载波相位、精密单点定位形成了优化的定位方法；2）提出了数字轨道地图辅助的多传感器组合定位方法，分别基于联邦滤波、紧组合策略、二分搜索和移动窗口联合的地图关键点搜索方法形成了改进的组合定位方法；3）提出了列车定位传感器量测级别的故障检测与自适应容错机制，通过“全局最优滤波估计”、“量测故障检测识别与分类”、“故障信息的分类自适应处理”集成形成了列车鲁棒定位方法。项目相关研究成果为将卫星导航系统引入列车定位与列车控制系统提供了方法支持及前期验证，对促进我国北斗卫星导航系统在轨道交通领域的应用及基于位置的轨道交通列车运行控制的上层应用发展具有积极意义。