This paper provides a systematic review of emerging control techniques used for railway Virtual Coupling (VC) studies. Train motion models are first reviewed, including model formulations and the force elements involved. Control objectives and typical design constraints are then elaborated. Next, the existing VC control techniques are surveyed and classified into five groups: consensus-based control, model prediction control, sliding mode control, machine learning-based control, and constraints-following control. Their advantages and disadvantages for VC applications are also discussed in detail. Furthermore, several future studies for achieving better controller development and implementation, respectively, are presented. The purposes of this survey are to help researchers to achieve a better systematic understanding regarding VC control, to spark more research into VC and to further speed-up the realization of this emerging technology in railway and other relevant fields such as road vehicles.
翻译:本文件系统地审查了铁路虚拟组合(VC)研究中使用的新兴控制技术;首先审查了培训运动模型,包括模型配方和所涉及的部队要素;然后详细阐述了控制目标和典型的设计限制;然后对现有VC控制技术进行了调查,并分为五组:基于共识的控制、模型预测控制、滑动模式控制、机器学习控制和制约控制;还详细讨论了这些技术对VC应用的利弊;此外,还介绍了今后为实现更好的控制开发和实施而开展的若干研究;调查的目的是帮助研究人员更系统地了解VC控制,激发对VC的更多研究,进一步加快在铁路和其他相关领域(如公路车辆)实现这种新兴技术。