Multi-Objective Real-Time Weighted Model Predictive Control for Car-Following

In this study,a multi-objective adaptive cruise control algorithm for complex traffic scenarios is developed. 4-DOF vehicle dynamics model taking into consideration of tightly coupled relationships between longitudinal and lateral dynamics is established that is then completely decoupled into a 1st order linearized longitudinal car-following model as well as a 1st order differential equation based lateral dynamics model by utilizing a state-feedback based disturbance decoupling strategy. Moreover,a linear parameter-varying method is employed to discretize the equations of the continuous lateral dynamics subsystem that makes it unnecessary to execute discretization p-times in each control cycle,thereby considerably decreasing the high computational complexity. Furthermore,to achieve dynamic coordination between longitudinal and lateral performance,a weight coefficient self-tuning strategy based on a model predictive control(MPC)framework considering multi-traffic scenarios fusion design(MPC-RW*)is suggested by which the weight coefficient for each sub-objective can be adaptively adjusted along with the change of traffic scenarios. The comparative simulations show that MPC-RW* can achieve better response during car-following. © 2020, Editorial Board of Journal of Tianjin University(Science and Technology). All right reserved.