Reconfigurable Model Predictive Control for Virtual Track Train Path-Tracking Considering Hinge Force

Virtual track train (VTT) is a new multi-articulated public transportation system that adopts rubber wheel and rail transit operating management mode. Reducing the hinge force while ensuring high path-tracking accuracy is a hot issue for the motion control of the distributed driving all axle steering virtual track train (DDAAS-VTT), which is very important to improve the train safety and service life. In this paper, the dynamic model of the DDAAS-VTT is established by a two-step method to realize the rapid expansion and reconfiguration according to the marshalling number. The hinge force is analyzed and expressed by the vehicle parameters and control inputs. A path-tracking strategy is proposed based on a hierarchical framework, which is mainly composed of two parts: a cooperative controller (CC) based on an improved model predictive control (MPC) algorithm and considering the hinge force, and single unit controllers (SUC) for the control allocation (CA) of wheel steering angles and torques. The co-simulation results show that the proposed control strategy can achieve high path-tracking accuracy, perfect curve passing performance, and robustness. The controller expansion and reconfiguration for different numbers of vehicles are easy to implement. Compared with other algorithms, the proposed method can effectively reduce the hinge force without increasing the computational press of the ECU.