A Fuzzy Logic-Based Model Predictive Control Framework for Risk-Aware Safety Control of Railway Virtual Coupling

Virtual Coupled Train Sets (VCTS) achieve cooperative operation through inter-train communication; however, under real-world operating conditions, multi-source operational risks pose significant challenges to coordination and safety. To address this issue, this paper proposes a multi-layer risk-aware reference generation and control framework for VCTS to enhance the operational safety of train formations. Within this framework, VCTS is abstracted as a leader-follower train structure, and the multi-source risks faced by the following train are categorized into leader-train system-level risk, environmental risk, and communication risk. Specifically, the leading train employs a Bilateral Cumulative Sum (Bi-CUSUM)-based module to predict its system-level risk and transmit it to the following train. The following train then integrates the leader's system-level risk, onboard-perceived environmental risk, and communication risk using a fuzzy logic approach based on the analytic hierarchy process (AHP), resulting in a comprehensive risk assessment index. Based on this index, a dynamically generated, risk-adaptive target speed serves as the reference for a receding-horizon Model Predictive Control (MPC) scheme, enabling safe and conservative longitudinal behavior under dynamic operating conditions. Simulation results demonstrate that the proposed framework is capable of providing safe and robust control performance under elevated risk conditions.

Keywords: Virtually Coupled Train Sets, Risk-Aware Control, Fuzzy Logic, AHP, Bi-CUSUM, MPC.