Studies On Lane Changing Algorithm Of Intelligent Vehicle On Highways

Due to the small traffic density and one-way driving on the same side,highways are one of the most suitable scenarios for the intelligent vehicle to achieve automatic driving.Therefore,the technology of intelligent vehicle on highways is always the hot topic in academic research.The lane change is one of the basic driving behaviors on highways,and it is also the main cause of traffic congestion and accidents.Therefore,this thesis relies on the key research and development plan of the Ministry of Science and Technology,“Key Technology Research and Demonstration Operation of Electric Autonomous Vehicle” to study the lane changing algorithm of intelligent vehicle on highways,main goal is to develop an algorithm that meet safety,legitimacy,comfort and rationality.Firstly,this thesis studies the behaviors of drivers on highways based on driving simulator and questionnaire,the triggering process of the driver’s lane change and the change law of the kinematics and dynamic parameters of the vehicle during the lane change process are analyzed.It is concluded that the dissatisfaction of space or speed is the main causes of discretionary lanechanging,the lane change time is 4~6s,and during the lane change,the maximum lateral acceleration is 0.2g,the maximum yaw angle is 5 degrees,the shape of lane change trajectory is like high-order polynomial.Secondly,based on above conclusions,this thesis designs a lane changing decision algorithm based on decision tree.The decision algorithm consists of two parts.The first part calculates the dissatisfaction degree of lane.When calculating the dissatisfaction degree of lane,it fully considers all kinds of scenarios.The second part selects the target land and issues instructions.When selecting the target lane,the lane speed advantage,legitimacy,safety are comprehensively considered.The lane changing decision algorithm will eventually give a binary command: lane change or no lane change.The lane change direction will also be indicated when the lane change command is issued.Due to most of the lane change trajectory planning algorithm is static.This thesis designs a dynamic trajectory planning algorithm based on adaptive potential field.Firstly,based on the potential field method,the algorithm establishes the driving area centered on the ego vehicle,and an adaptive factor is used in the traditional potential field function,which can automatically change the potential field distribution with the change of relative speed of the two vehicles.Secondly,the algorithm generates a series of trajectory clusters based on the fifth-order polynomial according to the current vehicle states,Finally,the algorithm selects the optimal trajectory consider safety and comfort.The trajectory planning algorithm performs the above steps in each cycle,so the lane changing trajectory can be dynamically planned.Forth,this thesis designs lane change trajectory tracking algorithm base on model predictive control.Firstly,the vehicle prediction model on the highways is established according to the application scenario of the algorithm,then the corresponding controller is designed according to the constrains in the lane changing process.Finally,the constrained optimization problem is transformed into QP problem.The algorithm not only consider the tracking accuracy but also the comfort.Finally,the lane change algorithm designed in this thesis is verified base on Simulink/CarSim and vehicle.The results verify the rationality and validity of the algorithm,and prove that the lane change algorithm designed in this thesis has certain robustness.