Vehicle Trajectory and Velocity Planning in Mixed Traffic Scene with Other Vehicles by Using Nonlinear Model Prediction Control and Dynamic Programming

Abstract

In this paper, we propose a vehicle trajectory and velocity planning method for autonomous driving. The proposed method includes a nonlinear model predictive control and a dynamic programing. The dynamic pro-graming part calculates an optimal path and a driving corridor to satisfy the safe zone took into account other vehicles traffic situation. The optimal path and the driving corridor are used for nonlinear model predictive control part as a target value and a constraint. The nonlinear model predictive control part calculate a vehicle trajectory and velocity simultaneously for a comfortable drive. To confirm the effectiveness of these proposed methods, numerical simulations are demonstrated. Simulation results show that the proposed method can calculate a humanlike driving behavior. For example the velocity planning results show slowing down before corner entrances and the trajectory planning results show the out-in-out shape to reduce the curvature of the vehicle trajectory.