Research On Autonomous Lane-Changing System Of Self-driving Vehicles

Intelligent Vehicles have been widely concerned by the public because of they have great potential in improving driving safety and road traffic efficiency.As one of the common driving behaviors,Lane-changing System has an important influence on traffic efficiency and safety.This article analyzes the Lane-changing System of Selfdriving Vehicles based on the highway scene.And this article elaborates four aspects: Vehicle Dynamics Modeling and Motion Control,the Decision?Motion Planning and Tracking of Lane-changing System.In the aspect of Vehicle Dynamics Modeling and Motion Control,firstly,the target vehicle is described as a-fourteen-DOFs vehicle dynamics model.Then,a layered motion control algorithm framework is designed,that is,the upper layer is the motion control algorithm,and the lower layer is the torque coordination allocation.Finally,simulation experiments have been designed to verify the effect of motion control.From the analysis of experimental data,the motion control algorithm has satisfied control effect.In terms of Lane-changing Decision,first of all,according to the analysis of the characteristics of Lane-changing behaviors,Selective Lane-changing behavior is determined to study in this paper that is,to get rid of the low-speed vehicle and obtain wider driving space and higher speed.Finally,Lane-changing Decision which satisfies the intention and feasibility of Lane-changing is designed.The Lane-changing Decision is obtained,the target lane information is also obtained.In the aspect of Lane-changing Movement Planning,firstly,the basic theory of Frenet coordinate system is introduced.Then,the relationship between cartesian coordinate system and Frenet coordinate system is derived.Finally,the highdimensional trajectory optimization problem is transformed into multiple onedimensional optimization problems based on Frenet coordinate system,reducing the complexity of the trajectory optimization problem.Then,the set of alternative trajectories was generated by using quantic-polynomials.Finally,the objective function is designed to achieve the screening of the optimal trajectory.In the aspect of Tracking,the vehicle dynamics model with transverse and longitudinal coupling is established firstly.Then,linear discretization is carried out based on the model,and vehicle stability constraints are introduced.Finally,MPC(Model Predictive Control)is converted to online problem solving.Through the analysis of the track following simulation experiment under the conditions of high and low adhesion road surface,the track following controller designed in this paper has a good track following effect,and can still ensure that the wheels are in the linear region under the limit working condition,to avoid the instability phenomenon of the vehicle under the limit working condition.Finally,combined with the real vehicle platform,the tests of straight lane keeping,bend lane keeping,left lane changing,right lane changing and other working conditions are designed.According to the analysis of test data,the self-lane changing system designed in this paper has better performance.