Research On Key Technologies Of Lane Change Control For Autonomous Vehicle Based On Safety Of The Intended Functionality

In recent years,autonomous vehicle technology has become the focus of the global automotive industry because it can effectively solve traffic safety problems caused by improper driving behavior,fatigue driving,negligent or illegal driving,and other human errors.Meanwhile,the rapid development of related technologies such as sensing,information and artificial intelligence provides a good hardware and software foundation for the practical application of autonomous driving technology.Both the government and enterprises are vigorously promoting the development and application of automotive automation technology.However,due to technological maturity and cost,autonomous driving technology is still in a stage of continuous development.However,due to the constraints of technology maturity and cost,autonomous vehicle technology is still in the stage of continuous development.In this dissertation,based on the safety of the intended functionality of autonomous vehicle,the theoretical and experimental research on the automatic lane change control are carried out,include the lane changing decision,the lane change path planning,the path tracking and control execution.This dissertation first reviews the research background and development history of the lane change technology for autonomous vehicle.Then the safety of the intended functionality and automatic lane change control technology of autonomous vehicle are introduced.And the research status of prediction-decision,the lane change path planning,the path tracking and control execution are summarized.In order to improve the safety of the lane changing decision of autonomous vehicles,after the surrounding vehicles are divided into related vehicles that only need to consider their longitudinal motion state and related vehicles that need to be considered in both vertical and horizontal motion states,the deep neural network and the lateral deviation judgment criteria are designed to judge the lateral motion behavior of associated vehicles.On the basis of considering the impact of lane-changing and lane-departure of the associated vehicles in the adjacent second lane on the safety of lane-changing,the safety situation of the adjacent lanes of the controlled vehicle is evaluated,and the lane-changing decision criterion is designed to realize the accurate lane-changing decision to reduce the safety of the intended functionality issue.According to the safety of the intended functionality issue that safety of lane change is difficult to ensure for different lane changing environments in most lane change studies,the lane change path is expanded into a restricted lane change driving space,a lane change control method based on vehicle driving safe boundary is proposed to solve this problem.Accounting for the lane change space and the collision time interval between the controlled vehicle and the surrounding associated vehicles,a method for dividing lane change state is designed.Secondly,the vehicle position safe boundary considering the vehicle-road position relationship and the vehicle safe rectangular area is established.Based on the driving stability limit and the requirements of yaw rate in different lane changing states,a safe boundary of the vehicle motion state corresponding to the lane changing states is established.The model predictive controller is designed to determine the desired front wheel angle to achieve safe,comfortable and stable lane change control,in which the vehicle position safety boundary and the vehicle driving safety boundary are used as the constraints of the vehicle lane changing state parameters,and the centerline of the target lane is used as Reference trajectory.To further improve the lane changing performance of autonomous vehicles,a path tracking method based on model predictive control with variable predictive horizon is proposed on the basis of analyzing the influence of the predictive horizon on the vehicle path tracking control performance.Based on the designed model predictive controller for path tracking,the response analysis of the path tracking control system under the different predictive horizons is carried out to clarify the influence of predictive horizon on path tracking.Then,taking the lateral offset,the steering frequency and the real-time of the control algorithm as comprehensive performance indexes,the particle swarm optimization algorithm is designed to realize the adaptive optimization for the predictive horizon.By selecting different optimal prediction time domains in different driving environments,the tracking accuracy of the vehicle is higher while the steering is more smoothly.To reduce the adverse effect of the functional insufficiency of the steering system on accuracy of path tracking,a cooperative control approach considering safety of the intended functionality based on automatic steering and differential braking is proposed for lane change path tracking.The proposed method adopts a hierarchical architecture consisting of a cooperative control layer and an execution control layer.In the cooperative control layer,an extension controller considering functional insufficiency of the steering system,tire force characteristics and vehicle driving stability is proposed to determine the weight coefficients of automatic steering and the differential braking,and a model predictive controller is designed to calculate the desired front wheel angle and additional yaw moment.In the execution control layer,the active disturbance rejection controller is designed to determine the assistance torque,and a braking force distribution module is used to determine the wheel cylinder pressure of the controlled wheels,so that the differential braking can reasonably compensate the automatic steering system,and reduce the safety of the intended functionality problems caused by the functional insufficiency of the steering system during the lane-changing path tracking.Based on Lab VIEW and Car Sim,the hardware-in-the-loop test benches established.The proposed lane-changing control method based on vehicle driving safe boundary,path tracking algorithm based on the MPC with variable predictive horizon,and the path tracking method based on cooperative control of automatic steering and differential braking are tested and verified.Based on a certain electric vehicle,multiple types of environmental sensing sensors,computing and processing platforms,automatic steering control systems,and active braking systems are used to build an autonomous vehicle test platform.The experimental results verify the effectiveness of the proposed lane changing decision,lane changing control method and path tracking method.