Motion Control And Decision Planning For Car Following And Lane Changing Of Autonomous Vehicle

In the context of informatization and intelligence in the automotive industry,the autonomous vehicle technology has gradually become the hot spot.This thesis is established in the research and study for the autonomous vehicle technology in level 4,aiming at two typical driving behavior,namely car following and lane changing,when the autonomous vehicle is driving on structured highways,the main content focuses on vehicle Longitudinal motion control,Lateral motion control,Driving behavior decision and lane-changing trajectory planning.The main research work and innovative achievements of this paper are as follows:(1)Aiming at the problem of the uncertainties of vehicle longitudinal dynamics model and actuator model,an adaptive cruise control strategy based on? synthesis control theory is proposed.The adaptive cruise control involves braking system and drive system,which all contain strong model uncertainty and are easily affected by external interference,these characteristics will significantly affect the dynamic performance of vehicle control system and then affect the stability of the control system.In this paper,a robust adaptive cruise control strategy is proposed:firstly,a three-state space model for car-following system is built,after that,in order to decouple the uncertain parameters and reduce the conservativeness of the controller design,the state-space model for the vehicle-following system is reformulated using the descriptor form,finally,a robust adaptive cruise control strategy is designed based on the ?-synthesis theory,which provides a new solution to improve the robustness of adaptive cruise control system.(2)Considering the influence of differential braking characteristics on lateral stability control,a multi-system cooperative control strategy of autonomous vehicles lateral motion control is proposed based on the seamless intervention mechanism of differential braking.For the conventional vehicle,the direct yaw moment in lateral tracking stability control is usually produced by differential braking,however,frequent braking will disturb the vehicle longitudinal dynamic and significantly affect the vehicle speed.In this paper,a cooperative control strategy of steering and braking is proposed based on seamless intervention mechanism of differential braking,that is,in the case of normal driving situation,only the front steering control system is adopted to realize the lateral stability control for path following of autonomous vehicle,and when the critical situation is detected,the direct yaw-moment control using differential braking is activated gradually and both steering and braking work together to ensure vehicle stability and safety.The control strategy adopts the hierarchical architecture: the upper layer designs the desired yaw rate generation strategy based on the vehicle lateral motion model and backstep theory;by considering the vehicle longitudinal velocity as a time varying parameter,the lower layer designs a cooperative control strategy of steering and braking based on the vehicle lateral dynamics model and LPV control theory,which is used to track the desired trajectory.The proposed control strategy can ensure the tracking accuracy and lateral stability while reducing the influence of differential braking for the vehicle longitudinal dynamics,which provides a new solution for the multi-system cooperative control of the vehicle motion control.(3)Aiming at the problem of sudden fault or failure for electronic hydraulic braking(EHB)system in the process of improving vehicle lateral dynamic stability by differential braking,a generalized fault model of EHB system is built and a fault tolerant control strategy based on robust fault-tolerant control theory are proposed.The EHB system consists of four independent brake wheel cylinder,which is a typical overdrive system.In this paper,firstly,a generalized fault model is built for the four typical fault types which may occur in the process of lateral stability control of the braking system,including no fault,loss-of-effectiveness fault,stuck-at-fixed level fault,and additive fault.After that,by considering the fault factors in the fault model as scheduling parameters,and based on the robust control theory,the fault-tolerant control strategy is designed for EHB system in the process of the lateral stability control.The fault-tolerant control strategy proposed in this paper can effectively deal with the four kinds of typical fault types mentioned above,that is,for automatic vehicle lateral stability control system,when one or more of the braking unit fault or failure occurs,the proposed control strategy can combine with the steering system and coordinate the braking ability of each wheel to ensure the vehicle travel stability,and realize the faulttolerant control for the braking system.(4)Considering the problem of driving situation-adaptive for the decision making and planning of autonomous vehicle in the driving scenarios of vehicle following and lane changing,an anthropomorphism driving behavior decision method and lane changing trajectory planning strategy are proposed based on Classification and Regression Tree(CART)and Generative Adversarial Networks(GAN)method.Vehicle following and lane changing behaviors are the most common traffic behaviors in the real driving process,vehicle following behavior is deeply affected by forward traffic flow,and the lane changing behavior involves more lanes and vehicles,and in the mixed traffic flow,the autonomous vehicle has strong dynamic under the influence of surrounding vehicles.And the human drivers can easily cope with a variety of dynamic and complex driving environments.In this paper,by studying and screening the information features that directly affect the results of driving behavior decision-making in the real driving process,a driving behavior decision-making model based on CART decision tree is proposed,and by analyzing the Gini index of each feature information,the relative ranking of the importance of each feature information for the driving behavior decision-making is obtained,finally,a driving behavior decision tree model with clear classification rules and relatively compact structure is constructed.After that,by screening lane-changing trajectory of expert demonstrations in real human driving data set,a lane-changing trajectory planning strategy based on the GAN method is proposed,and through the neural network game training,a neural network model that can generate the lane changing trajectory closing to the expert demonstrations is constructed.Through the combination of driving behavior decision and lane change trajectory planning,not only the reasonable lane change timing can be obtained,but also the lane change trajectory,which is consistent with the human driving characteristics,can be obtained,which improves the ability of autonomous vehicles to adapt the dynamic traffic environment.(5)According to the proposed motion control strategy and decision planning method,the automatic driving system of vehicle following and lane changing is constructed.Based on the real traffic flow information and the hardware-in-theloop experiment platform,the effectiveness of the proposed control strategy and algorithm is verified by comprehensive experiments.The results show that,based on the lane-changing trajectory planning model and lateral trajectory tracking strategy designed in this paper,autonomous vehicles can achieve stable lane changing and ensure high lane-changing trajectory tracking accuracy under different driving conditions.And for the vehicle following behavior under different driving conditions,based on the longitudinal following control strategy designed in this paper,the autonomous vehicle can achieve fast and stable vehicle following behavior.In conclusion,aiming at the application requirements of the automatic driving technology,this paper focuses on the motion control,decision-making and planning of the vehicle following and lane changing,and puts forward the vehicle Lateral motion control,Longitudinal motion control,driving behavior decision and lane-changing trajectory planning solutions,and a complete vehicle autonomous driving system of vehicle following and lane changing is constructed,which realizes the multi-system cooperative work for the autonomous vehicle following and lane changing,and improves the vehicle travel stability and safety.