Trajectory Planning Based On Edge Computing For Vehicle Autonomous Driving

As a current research hotspot in the field of traffic,autonomous driving technology can effectively improve driving convenience and alleviate traffic safety.It is one of the critical technologies of automatic driving that the vehicle generates stable and safe driving trajectories.The vehicle judges the surrounding driving environment through environmental perception technology,and generates the optimal trajectory suitable for the current environment,so as to control the vehicle to avoid obstacles.In order to realize autonomous driving of a vehicle,the vehicle will generate a large amount of computational data.In the traditional cloud computing mode,the vehicle transmits the data to the cloud computing center,and the processing results are returned after calculation processing,which may increase the system delay.For high-speed self-driving cars,at least millisecond response time is required,and cloud computing centers are faced with huge amounts of data.Once the computing delay occurs,it will cause serious consequences.Therefore,edge computing is required,for which all the data generated by the vehicle is not necessary to be uploaded to the cloud for processing,and is stored and calculated at edge nodes.It can meet the real-time requirements of autonomous vehicles.In this thesis,a vehicle trajectory planner for lane-changing overtaking situations is designed,which is suitable for the high-speed driving of vehicles.According to the vehicle’s driving states and road environment,it can perceive and detect the vehicle in the environment ahead,and generate a lane-changing and obstacle-avoiding safe trajectories to ensure the safety of autonomous vehicles.In addition,based on the edge computing architecture,this thesis builds a vehicle trajectory planning platform based on edge computing.The research contents and results of this thesis are described as follows:(1)A vehicle driving environment model is established based on the artificial potential field method.The characteristics of the vehicle driving environment at high speeds are analyzed,and the artificial potential field model of the vehicle driving environment is established.At the same time,according to the characteristics of the road environment in the case of overtaking and obstacle avoidance,a feasible area for vehicle overtaking under the overtaking situation is proposed,and a potential field function is established,which improves and enriches the vehicle driving environment models in case of overtaking and obstacle avoidance.(2)A trajectory planner is proposed based on MPC and APF.Firstly,a point-mass model of vehicle dynamics is established to describe the motion states of the vehicle.According to the MPC rolling optimization mechanism and the established APF road potential field model,the trajectory planning problem of vehicle operation is transformed into an optimization problem in the prediction horizon.Secondly,an overtaking trajectory optimization strategy is designed to alleviate the problem of trajectory fluctuation caused by the objective function is calculated to switch the centerline of the left and right lanes in the overtaking procedure of changing lanes.Finally,the genetic algorithm is used to optimize the parameters of each potential field weight value in the objective function in the current environment,and the vehicle trajectory planner generates the optimal driving trajectory according to the current environment and vehicle states based on the obtained optimization weight parameters..(3)A vehicle trajectory planning platform is established based on edge computing.The overall design is carried out according to the performance characteristics of platform development,including the vehicle edge device and the edge computing server.The platform consist of human-computer interaction interface design,vehicle positioning and map display module,communication interface module,data analysis module,data storage module and information transmission module..(4)Simulation experiments and functional tests are carried out on the designed vehicle trajectory planner and the trajectory planning platform based on edge computing.Three longitudinal simulation tests are designed to verify the optimal performance and effective performance of vehicle trajectory planning,and lateral comparison tests at different relative speeds are designed to verify the applicable scope of the vehicle trajectory planner.In addition,the functional test of the edge computing-based vehicle trajectory planning platform is carried out to verify the feasibility of various modules.70 Figures,15 Tables,58 References.