| With the development of a new round of science and technology,the living standards of the people have been greatly improved,and the number of cars owned has shown a trend of increasing year by year.At the same time,traffic congestion and traffic safety issues are also becoming increasingly serious.The proposal of autonomous driving technology is an effective solution to reduce the occurrence of these two problems,and the research on autonomous driving technology has become the main development direction in the field of automotive industry in recent years.The intelligent vehicle with autonomous driving function is a very large and complex system,in which the trajectory planning module and trajectory tracking control module are important components.This paper takes the intelligent vehicle in the structured road cruise scenario as the research object.The trajectory planning problem is decomposed into path planning and velocity planning by Frenet coordinate system.The trajectory tracking control is divided into horizontal control and longitudinal control.The research content of this paper is as follows:(1)The coordinate systems commonly used in automatic driving algorithm research are introduced,and the transformation relationship between absolute coordinate system and Frenet coordinate system is deduced by vector method.Then the selection principle of projection point in Frenet coordinate system is analyzed and studied.Finally,Matlab simulation platform is used to verify the transformation relationship between coordinates.(2)The path planning algorithm is designed based on Frenet coordinate system.Firstly,a smooth guide line is generated by constructing a guide line module as a road guide line in Frenet coordinate system.Then the traditional artificial potential field is improved to establish the dangerous potential field of road and obstacle respectively and generate a rough path by combining the quintic polynomial.Finally,quadratic programming algorithm is used to optimize the rough path and generate a fine path.(3)Equipping the generated fine path with speed information.Firstly,ST coordinate system is established and sampled according to the generated fine path and the predicted trajectory of moving obstacles,and then a rough velocity curve is found in ST coordinate system using the dynamic programming algorithm.At the same time,in order to improve the search efficiency of the dynamic programming algorithm,the generated path is sampled unevenly and the SV coordinate system is established.Finally,the solution space of the velocity optimization problem is constructed according to the rough velocity curve,and the final velocity curve is obtained by solving the velocity optimization problem in the solution space.(4)The path curve and the velocity curve are synthesized into a trajectory as the input of the trajectory tracking control module.In order to reduce the difficulty of the control,the Frenet coordinate system is used in this paper to decouple the trajectory tracking control into lateral and longitudinal control.For the lateral control,the two-degree-of-freedom lateral dynamics model of the vehicle was established and the error model of the vehicle was derived based on Frenet coordinate system.Then the lateral LQR tracking controller is designed according to the vehicle error model.Finally,feedforward compensation and error compensation are used to reduce the trajectory tracking error.For longitudinal control,in order to reduce the modeling difficulty of the vehicle power system,the throttle/brake calibration meter is used instead of the vehicle power system modeling.A layered PID controller is designed to control the throttle opening or brake pressure of the vehicle to achieve longitudinal displacement and speed tracking control. |
PDF Full Text Request