Constrained Model Predictive Control for Nonholonomic Vehicle Regulation Problem

A model predictive control architecture based on discrete time nonlinear car model is derived to solve regulation(“parking”) problem. Parameters of the proposed controller are chosen by considering terminal state constraints. This setup combined with terminal state penalty in the cost function could assure control stability. The generated trajectory satisfies minimum curvature requirements and actuator saturations of the vehicle are considered in controller design. Obstacle avoidance is realized by considering distance constraints in the openloop optimization process. Simulation results are given to illustrate the feasibility of the proposed control architecture.