Research On Autonomous Lane Changing Decision-Making And Motion Planning Technology Of Unmanned Vehicles Based On Driver Behavior Characteristics

Driven by the rapid development of computer technology and artificial intelligence technology,unmanned driving technology is setting off a global boom.Many technology companies,OEM and universities are actively exploring unmanned driving technology.Unmanned driving technology can not only bring great convenience to human’s daily life,but also provide perfect solutions in terms of ecological protection,safe transportation,and alleviation of urban traffic pressure.The realization of unmanned driving requires the coordinated development of sensing technology,positioning technology,fusion prediction technology,decision-making and planning technology,and wire control technology.In order to realize unmanned driving of vehicle,an important key technology that needs to be solved is autonomous lane changing technology,which enables unmanned vehicles to make correct lane changing decisions in specific driving scenarios and carry out reasonable motion planning to ensure that unmanned vehicles can make correct lane changing decisions,which make the lane-changing safe and reasonable.This article are based on the National Natural Science Foundation of China project "Research on the control mechanism and evaluation method of a new type of steer-by-wire system based on driver characteristics"(No.51575223)and the Science and Technology Project of Jilin Province Education Department "Distributed electric vehicle dynamic and collaborative control based on chassis-by-wire system"(No.JJKH20200963KJ).Based on the logic characteristics of real drivers’ behavioral decision-making when changing lanes in various driving scenarios,design a decisionmaking method for unmanned vehicle lane-changing behavior,and study the unmanned vehicle lane-changing motion planning method based on the real driver’s lane-changing trajectory.The specific research content of this article is as follows:Design of Decision System for Autonomous Lane Changing Behavior of Unmanned Vehicles:This article first analyzes the relevant parameters of the driver in the course of the normal lane change,and uses these parameters to study the driver’s lane change scene and behavior characteristics,to obtain the reference parameters of the real driver during and after the lane change.Through the record of the decision-making process of multiple real drivers who simulated driving in a passenger car driving simulator,the characteristics of the driver’s psychological activity when changing lanes are summarized,and the pavement for subsequent behavioral decision-making system research is made.Secondly,through the establishment of the kinematics model of the unmanned vehicle and the driving environment model of the unmanned vehicle,the behavior decision system is designed to consider the constraints of the entire vehicle kinematics model and the driving environment.Then,in the design of the unmanned vehicle’s autonomous vehicle lane-changing behavior decision system,four types of unmanned vehicle lane-changing intentions are established based on the driver’s lane-changing behavior characteristics,which are determined by the left and right priority design of the various lane-changing intentions of unmanned vehicles.In the process of unmanned vehicles preparing to change lanes and executing lane changes,the right-of-way characteristics of unmanned vehicles and traffic vehicles are added.The design of the right-of-way feature ensures the safety of unmanned vehicles when changing lanes,and at the same time prevents other vehicles from inserting into the target.The design of this feature conforms to the driver’s driving behavior characteristics.Finally,by designing an autonomous lane-changing finite state machine,the various states of the unmanned vehicle in the lane-changing process are designed to realize a reasonable transition between various states and complete the decision-making level of the unmanned vehicle’s autonomous lanechanging behavior.The simulation verification shows that the design method conforms to the driving behavior characteristics of real drivers.Research on Autonomous Lane Changing Motion Planning Method of Unmanned Vehicles:At the unmanned vehicle lane-changing motion planning layer,lane-changing trajectory planning is mainly based on the driver’s visual characteristics.First,through the analysis of the driver’s visual characteristics,an optimal design method for the perception layer is proposed.According to the different driving tasks of the driver,the visual area grid in the field of view is designed with different degrees of cost.Then by modeling the body dynamics and tire dynamics of the unmanned vehicle,the dynamic constraints of the unmanned vehicle during the lane changing process are proposed,so that the lane changing trajectory planned by the motion planning layer meets the requirements of vehicle dynamics.Finally,based on the Bezier curve model,the trajectory planning of the unmanned vehicle’s lane changing under constant speed driving conditions and variable speed driving conditions is carried out,and the unmanned vehicle’s constant speed lane changing trajectory planning system is carried out based on different risk coefficient requirements and trajectory curvature requirements.Based on different lane-changing time requirements and lane-changing acceleration requirements,the unmanned vehicle’s variable-speed lane-changing trajectory planning system is designed.The unmanned vehicle lane-changing motion planning system designed by the simulation platform is verified.Simulation test verification based on passenger car driving simulator:At the end of this paper,the above research content is simulated and verified based on the passenger car driving simulator in the laboratory.First,set up an unmanned vehicle lane changing scene in the driving modeler.Based on driving experience and the priority value and risk factor of the unmanned vehicle’s lane changing process in the set scene,the designed unmanned vehicle lane changing behavior can be determined.The final lane-changing intention selection of the unmanned vehicle is determined by calculating the priority of each lane-changing intention during the lane-changing process,and the risk factor and driving experience are used to determine that the current lane-changing choice of the unmanned vehicle is reasonable and effective.Based on the unmanned vehicle dynamics constraints in Chapter 3,the dynamic parameters of the unmanned vehicle’s lane change process are judged,and the designed lane-changing trajectory is determined to have high lanechanging efficiency and can meet the unmanned vehicle dynamics constraints,and the lateral acceleration of lane change is small,which can provide passengers with better riding comfort.