Curvilinear Path Following of an Autonomous Vehicle by Predictive Control Model

This paper presents a model predictive control algorithm for path following of an autonomous vehicle in a curvilinear path. To define vehicles’, motion a nonlinear bicycle model of vehicle with 6 degrees of freedom is employed. Some kinematic constraints imposed to the state and control variables, as well as constraints on vehicle acceleration and braking deceleration for desired path following of the vehicle. Then, an appropriate prediction horizon and a quadratic optimization model is formulated and employed for minimizing the path following error and control variable’s variation of the vehicle. In the optimization computation process suitable weight coefficient for all state and control variables are considered. Finally, a simulation study is performed to analyze the capability and efficiency of the presented algorithm for following the desired path. The simulation results show that the presented path following algorithm is acceptably able to follow the desired path with acceptable accuracy and desirable speed.