Time-suboptimal predictive control of four-in-wheel driven electric vehicles

The paper deals with the approximately time optimal control of four in-wheel-driven (4WD) electric cars in a test path under state and input constraints with initial perturbations. The path is divided into sections allowing that path information for the actual section appears in real time based on sensor fusion. For each section a separate optimum control problem is solved in a receding horizon predictive control (RHPC) fashion using the single-track model (2WD) of the vehicle. The problem is given as a dynamic nonlinear optimal control problem (DNOCP) and solved by reformulating it to a static nonlinear program (NLP) using discretization and direct multiple shooting methods. A novel method is presented to convert the RHPC optimal solution to the optimal control of 4WD cars. The conversion assures similar motion of the CoG points of both models and optimal distribution of the longitudinal wheel forces. For closed loop control of 4WD vehicle a discrete time model predictive control (MPC) is proposed which uses the optimal reference signals and the distributed wheel forces and optimizes the perturbations with analytically solvable end constraints.