A Data-Driven Distributed Iterative Learning Control Approach for High-Speed Trains

The safe and efficient operation of high-speed trains is a critical challenge, particularly in maintaining safe relative distances between adjacent trains in a set. Collaborative control among trains, particularly within multi-agent systems, has attracted significant research interest to address this issue. This paper extends existing single-agent iterative models to a multi-agent framework, proposing a data-driven approach to solve the iterative and asymptotic tracking problem for multi-agent train sets under unknown time-varying parameters and nonlinear disturbances. The proposed method incorporates a multi-agent communication network for state sharing, aligning more closely with real-world train operations. Additionally, an iterative dynamic linearization technique is employed, which requires only input-output information during iterative operations, eliminating the need for detailed model information. The effectiveness of the approach is demonstrated through simulations, highlighting its potential for enhancing the collaborative control and operational efficiency of high-speed train sets.