Research On The Lane Changing Control Of Highway Intelligent Autonomous Vehicle

Lane changing control of unmanned vehicle is a complex vehicle behavior,involves key links of intelligent vehicle,such as perception,decision making,control execution and so on,which is one of the key technologies for the realization of automatic driving.Lane changing behavior of vehicles involves several lanes,which would cause interference to other vehicles in traffic environment,affect the traffic efficiency,and even cause traffic accidents.Therefore,in-depth research on lane changing control of unmanned vehicle to achieve safe and efficient lane change has important practical significance for the realization of automatic driving technology and the improvement of traffic environment.Based on expressway driving environment,this thesis has carried out an in-depth study on the various stages of vehicle lane changing control.First of all,aiming at the emergence of lane changing intention,a driver's dissatisfaction calculation method based on the expected speed and the front vehicle type is proposed,and the target lane of lane change is selected considering the degree of congestion of the left and right lane.For the lane changing feasibility decision,a lane changing decision-making model is built based on fuzzy logic,and the distances between main vehicle with the front vehicle in current line and the front and rear vehicle in target line are selected as the model inputs.And the membership function parameters of the model inputs are calibrated using genetic algorithm in the condition of front car and front heavy vehicle respectively based on lane changing data extracted from NGSIM(Next Generation Simulation)expressway database.The comparative analysis result shows that when the front car is a heavy vehicle,in order to ensure driving safety,driver execute lane change operation when the distance between the front vehicle and the main vehicle is relatively big and the distances between the front and rare vehicles in target lane and main vehicle is relatively small.It shows that the difference of the type of front vehicle will influent driver's decision on executing lane change.Secondly,for lane changing optimal trajectory planning,considering the physical constraints of vehicle track,namely the curvatures of starting point and end point of lane changing trajectory should be zero to ensure that the vehicle runs straight along the lane,the vehicle trajectory curvature should be continuous without mutation and lateral acceleration should be small in the process of lane change to ensure passenger to be comfort,this thesis puts forward the comprehensive weighted lane change trajectory based on isokinetic trajectory and sinusoidal trajectory,and it can be adjusted according to different weights of lane change time and lateral acceleration.This trajectory is compared with the other two optimized lane changing trajectories from the aspect of trajectory's curvature,lateral velocity and lateral acceleration,which proves the superior performance of this trajectory.Finally,the control effect of the vehicle tracking lane change trajectory planned before is verified using model predictive control.A three degree of freedom dynamic model of vehicle based on simplified tire model is established,and a model predictive controller for vehicle trajectory tracking control is designed.The joint simulation model of Carsim and Simulink is built to simulate the control effect under different vehicle speed.The results show that the controller has good effect on the tracking control of vehicle lane changing trajectory.The lane changing control method of intelligent vehicle studied in this thesis is of theoretical value and practical application value for consummating the research of lane changing control and achieving vehicle automatic driving.