Autonomous vehicle velocity and steering control through nonlinear model predictive control scheme

To apply model predictive control (MPC) to autonomous vehicle, certain compromise has to be made. To avoid this, we propose a nonlinear MPC (NMPC) controller, controlling vehicle velocity and steering simultaneously. The optimization solver is based on genetic algorithms, which provides a flexible structure to design the cost function and constraints in a more accurate and straightforward way. On-field test showed the NMPC can control the vehicle to follow the road accurately and consistently. Moreover, passengers' safety and comfort were well taken care of as both the vehicle movement and steering accelerations are confined explicitly in the NMPC constraints.