• CN: 11-2187/TH
  • ISSN: 0577-6686

Journal of Mechanical Engineering ›› 2026, Vol. 62 ›› Issue (16): 228-252.doi: 10.3901/JME.260730

Previous Articles    

Advances in Dynamic Control Technology for Train Autonomous Operation

LING Liang1, LI Taotao1, GAO Feng2, LI Hongwei3, CHEN Zhe4, LIU Jianxin1, HOU Ye5, WANG Xiyun6, ZHAI Wanming1   

  1. 1. State Key Laboratory of Rail Transit Vehicle System, Southwest Jiaotong University, Chengdu 610031;
    2. State Key Locomotive for Traction and Control System of EMU and Locomotive, China Academy of Railway Sciences Corporation Limited, Beijing 100081;
    3. CRRC Changchun Railway Vehicle Corporation, Changchun 130011;
    4. CRRC Zhuzhou Electric Locomotive Co., Ltd., Zhuzhou 412001;
    5. CHN Energy Shuohuang Railway Development Co., Ltd., Suning 062350;
    6. CHN Energy Baoshen Railway Co., Ltd., Baotou 014000
  • Received:2025-08-17 Revised:2026-02-06 Published:2026-10-14

Abstract: Automatic train operation(ATO), widely applied the rail transit domain, is systematically reviewed in this paper. The fundamental principles and system architecture of ATO are outlined, and the characteristics and application scenarios of different automation grades are analyzed. The status of two pivotal technologies in ATO, namely speed trajectory optimization and speed tracking control, is examined both domestically and internationally. The technical features and engineering practices of ATO in urban rail transit and high-speed trains are summarized. The unique challenges in the ATO of heavy-haul trains are elaborated, including safety issues related to longitudinal impacts, significant system nonlinearity coupled with large time delays, and operations on long downhill sections. Corresponding countermeasures and technical approaches are also presented. The development direction of ATO is prospected from aspects of global environmental perception and decision-making control. It is demonstrated that, although significant progress has been made in dynamic train operation for ATO, the integration of environmental perception data into critical modules such as speed trajectory optimization and dynamic safety control remains insufficient. Moreover, the collaborative mechanism across the full“perception-decision-control”chain is still inadequate, resulting in limited adaptability under complex operational scenarios. In the context of heavy-haul train ATO, speed trajectory optimization algorithms and speed tracking control algorithms are not yet tightly coupled with refined train dynamics models, making it difficult for control strategies to meet safety requirements under complex operating conditions. In the future, refined models for train operation safety control should be further developed, encompassing train system dynamics, dynamic characteristics of traction and braking systems, and train state perception. Data-driven technologies should be employed to conduct model order reduction and online calibration, thereby accurately describing the nonlinear dynamic behavior of trains under complex operating conditions. A global and accurate environmental cognition system should be constructed to improve the level of perception technology and enhance the safe operation capability of the system under complex operation diagrams and changing environments. The in-depth collaboration and autonomous evolution of perception, decision-making and control should be promoted, and deep reinforcement learning and multi-objective optimization algorithms should be introduced to realize the accurate deduction of train operation state and the dynamic generation of optimal driving strategies, to advance the digital and intelligent transformation of rail transit.

Key words: automatic train operation, dynamic control, high-speed train, heavy-haul train, urban rail transit, speed trajectory optimization, safety control, environmental perception

CLC Number: