Multi-Scene Trajectory Planning Of Intelligent Vehicle Based On Frenet Coordinate System

With the improvement of the level of science and technology,intelligent vehicle,as a scientific and technological innovation achievement integrating automatic control technology,modern sensing technology,artificial intelligence and other technologies,is an important strategic direction in the development process of vehicles,and represents the development degree of the industry to a certain extent.As an important part of the behavior decision module of intelligent vehicle,trajectory planning can reflect the technical level of intelligent vehicle to a certain extent.In this paper,the trajectory planning model was designed and optimized based on Frenet coordinate system.In order to select the optimal trajectory under comprehensive conditions,the trajectory quality evaluation function was improved.Through horizontal and longitudinal decoupling and coupling analysis,the evaluation conditions under different vehicle running conditions were analyzed and the simulation analysis was completed.Firstly,the advantages of Frenet coordinate system in driving behavior calculation are compared and analyzed,and the conversion formula between Cartesian coordinate system and Frenet coordinate system is completed.Then,trajectory planning was decoupled into horizontal and vertical dimensions based on Frenet coordinate system.Horizontal and longitudinal trajectory planning models of the intelligent vehicle were established respectively.According to the information of the initial and final motion states of the vehicle,the corresponding trajectory sets of horizontal and longitudinal motion were generated,and the horizontal and longitudinal trajectories were reconstructed into complete trajectories.Again from vertical and horizontal two dimensions will be intelligent vehicle travel into high speed and low speed on the horizontal,vertical to the chi,overtaking a lane change,speed,parking six scenarios,and further coupled lateral and longitudinal directions of trajectory planning,the design of highway entrance ramp scene,T intersection and intelligent vehicle trajectory planning around the island road scene.Based on vehicle kinematics model,OBB method of acceleration Jerk and vehicle collision avoidance detection,trajectory quality evaluation functions of different scenes were determined.Finally,a three-lane road simulation environment is built,and random environmental vehicles are configured to meet the more changeable road traffic conditions.Through the configuration of different initial states,the intelligent vehicle completes the complete path planning on the road section,obtains the safe,comfortable and efficient movement trajectory and completes the visualization.The trajectory planning model based on Frenet coordinate system proposed in this paper can simplify the calculation process of trajectory planning for intelligent vehicles.The coordinate system transformation method can not only improve the calculation efficiency,but also reduce the influence of road curvature on the trajectory planning results,making the calculated trajectory safer and more comfortable.At the same time,the model accords with the characteristics of vehicle kinematics,which can improve the accuracy of the generated trajectory while reducing unnecessary computation,and can be applied to more intelligent vehicle running scenarios.