Model Free Adaptive Predictive Iterative Learning Cooperative Control of Multiple High-Speed Trains Operation With Disturbances

For cooperative control of multiple high-speed trains (MHSTs), a new model free adaptive predictive iterative learning cooperative control (MFAPILCC) method is proposed in this paper. Firstly, the unknown nonlinearity of the MHSTs is handled with the full form dynamic linearization (FFDL) technology along the iteration domain. Then with the predictive mechanism, the unknown bounded time-iteration-varying external disturbances are estimated and predicted, and also be incorporated into the controller design, which achieves better disturbance rejection and system stability. Specially, under the proposed MFAPILCC method, apart from the convergence of tracking error of system could be guranteed, the modeling error of system which practically makes a difference in the cooperative tracking performance of MHSTs also could be guranteed to converge to zero. The convergence of the modeling error and speed tracking error of the system under the constructed method could be guranteed through rigorous theoretical derivation, while the distance between adjacent trains is always maintained within a safety range. Finally, a simulation on the Chinese Railway High-Speed Train (CRH-3) operation control system is supplied to validate the efficacy and superiority of the proposed MFAPILCC method for cooperative control of MHSTs. It is obvious that the theoretical justification and simulation effects either reflect efficacy of the proposed MFAPILCC method. Note to Practitioners—In engineering practice, high-speed train (HST) operates in a repeatable and complex environment, and is susceptible to various external disturbances, such as weather, railway lines, and passenger load, etc. Besides, since the speed of HST gradually gets faster, it always needs to keep a safe distance from adjacent trains to ensure safe operation. Based on the above practical engineering considerations, the motivation of this paper is to design a model free adaptive predictive iterative learning cooperative control (MFAPILCC) method for unknown nonlinear and nonaffine Multiple high-speed trains (MHSTs) under time-iteration-varying external disturbances. The historical operation data of MHSTs is used to describe the complex dynamic mapping relationship between input and output of the train and the total data usage can be determined by engineering applicators. Then based on this description, a corresponding predictive iterative learning control method is constructed. Particularly, the proposed control method constructs a predictive mechanism for prediction and compensation of time-iteration-varying external disturbances with the aim of ensuring safe and punctual operation of trains in complex operation environment. Furthermore, the operation information of adjacent trains is introduced into controller design that assures adjacent trains always keep a safety distance from each other. The convergence of the proposed MFAPILCC method is theoretically analyzed by rigorous derivation. Simulation effects on the CRH-3 high-speed trains operation system further indicate the validity of the proposed method. Actually, the proposed control method can be applied to cooperative control tasks for not only MHSTs but also other practical systems since only I/O data information of the system is required for the design of controller.