Adaptive Step-Length Model Predictive Control for Vehicle Virtual Track Keeping on Superhighways

To ensure the track-keeping performance of superhighway driving and improve driving safety, a virtual track-keeping control algorithm is proposed for the vehicle yaw dynamics model based on the discrete effect of Variable Step-length Model Predictive Control (VSMPC). The transverse swing angular velocity-lateral velocity phase plane is employed as the stability criterion for the virtual track environment and front wheel deflection angle constraints. This paper verifies the applicability of the vehicle traveling on superhighway straight lines, circular curves, and combined road sections from the aspects of real-time and accuracy, which demonstrates the effect and stability of the virtual track-keeping control algorithm under the conditions of different vehicle speeds. The results show that the virtual track-keeping control algorithm based on VSMPC can restrict the vehicle with a speed up to 180km/h to the virtual track and a lateral offset within 4cm, thereby maintaining well-driving stability. The VSMPC control strategy effectively avoids the problem of poor model fidelity caused by high vehicle speed and effectively balances the computational efficiency and accuracy of the vehicle under the state of super-high-speed driving.