Trajectory planning of autonomous vehicle based on Frenét frame system

Abstract

With the development of autonomous driving technology, the situations handled by autonomous vehicles are becoming more and more complex. Local path planning is one of the important technologies of intelligent vehicles in the face of complex situations. To integrate more factors into consideration for different scenarios, an optimal trajectory generation method based on Frenét frame system is presented in this paper for trajectory planning of autonomous vehicles. In the optimal trajectory generation method, path is divided into multiple segments along the reference line and the vehicle state is sampled in each path segment. The cost function to evaluate each sampled path point and the vehicle state is established by considering the trajectory smoothness, comfort, energy consumption and the deviation degree of the reference route, and the parameters of the cost function are adjusted according to the different influence degree of factors through different driving modes of the vehicle. The function of obstacle avoidance of the trajectory planning is realized by deleting the trajectory that is less than the safe distance from the obstacles. Simulation verification shows that this method can complete trajectory planning under static obstacles and dynamic obstacles to realize the functions of obstacles avoidance and overtaking.