Research on path planning for autonomous vehicle based on Frenet system

Abstract

Autonomous vehicles drive according to the expected trajectory to achieve horizontal and longitudinal control of the vehicle. Therefore, the research on path planning is of great significance. This paper mainly analyzes the road behavior of the car, studies the five-fold polynomial algorithm model, simulates the path trajectory of different behavior, and analyzes the rate of change of lateral and vertical velocity and lateral and vertical acceleration under different behaviors, to get the car prediction time. Using the optimal control theory, the linear space expression is established to study the variation of transverse velocity, transverse velocity and transverse deflection angle, and the expression of the system to achieve the steady-state response is given. The simulation results are compared and analyzed to find out the optimal path of different scenes and different driving behaviors, which provides theoretical basis for real vehicle debugging.