Research On Authority Transfer Between Driver And Assisted Driving System In Human-machine Co-driving System

With the development of the automobile power strategy,vehicle intelligence has gradually become an inevitable trend in the development of the whole automobile industry.Therefore,the development of autonomous vehicles and semi-autonomous vehicles is now becoming a research hotspot focused on by emerging and traditional car manufacturers.Due to the complexity,dynamics of the traffic scene and the limitations of intelligent vehicle driving technology,fully autonomous driving is difficult to achieve soon.Compared with fully autonomous driving,human-machine co-driving semi-autonomous driving in specific scenarios has been widely applied in commercial applications and has broader application prospects.Therefore,human-machine co-driving semi-autonomous driving in specific scenarios has become a research focus in the current automotive industry.Nowadays,the human-machine co-driving semi-autonomous driving system is designed for specific scenarios,the transfer of vehicle control authority between the driver and the controller is unavoidable,and the common authority transfer method process is stepwise or linear transfer.Different drivers have different driving characteristics,so how to consider the driver’s characteristics during the transfer of authority and the controller-led driving process,and how to realize a smooth transfer of authority while satisfying different human driver preferences becomes an important question in semi-autonomous driving.In this paper,a driver-vehicle-road(DVR)system based on the 2-DOF vehicle dynamics model,road model,and two-point visual preview driver model is established.Consider two types of authority transfer requirements: controller-led curve driving and driver-led curve driving,path tracking controllers considering driver characteristics,ideal driver characteristics,and uncertainty factors are established based on the proposed DVR system,respectively.According to different application scenarios,the authority transfer strategies are designed based on spline curve and second-order dynamic system respectively to help drivers realize the smooth authority transfer process,improve the path tracking accuracy and vehicle ride comfort in the process of vehicle authority transfer,and greatly reduce the driver’s physiological and psychological load.It has both theoretical and practical value in the research of semi-automatic driving.The main research contents of this paper are as follows:(1)Controller-led vehicle control authority transfer.A personalized path tracking controller based on the linear time-varying model predictive control method and a personalized flexible authority transfer strategy based on the spline curve method are proposed.A driver-vehicle-road system is proposed combined with a 2-DOF vehicle dynamics model,road model,and two-point visual preview driver model.This personalized path tracking controller links controller parameters to driver characteristics to achieve a personalized path tracking controller design.The personalized flexible authority transfer strategy considers driver characteristic parameters to adjust the authority transfer time and process.The effectiveness of the proposed authority transfer strategy and path tracking controller is verified through a large curvature curve path simulation.(2)Driver-led vehicle control authority transfer.A personalized robust path tracking controller considering uncertainty factors based on the linear matrix inequalities method and a second-order dynamic authority transfer strategy are proposed.Based on the proposed DVR system,the robust path tracking controller considers an ideal driver model,the uncertainty of vehicle longitudinal speed,and vehicle control authority.A pole configuration approach is also applied to improve the controller’s dynamic performance.By analogy with the mass-spring-damping system,the authority transfer process is dynamically adjusted for vehicle longitudinal speed,road curvature radius,and driver steering behavior operation behavior,making the authority transfer process more consistent with the actual driving conditions of the driver.The effectiveness and applicability of the proposed controller and stagey are verified by simulating with different curvature bend paths.(3)Driver-in-the-loop driving simulator experiment.The hardware platform consists of a desktop,Logitech G29,and a triple-screen monitor.The software platform consists of Matlab/Simulink and Prescan.The experiments were conducted with different driving experience volunteers under the overall framework of the human-machine cooperative control authority transfer framework.Firstly,for the novice drivers taking the controller-led vehicle control authority on the curve as the test scenario,the proposed personalized MPC path-tracking controller and personalized flexible authority transfer strategy are used to compare with the traditional authority transfer strategy,the effectiveness of the proposed method is verified.Subsequently,for skilled drivers taking the driver-led curve driving vehicle control authority as the test scenario.The practicability and applicability of the proposed personalized robust path tracking controller and the second-order dynamic authority transfer strategy are verified by comparing it with the traditional transfer scheme and verifying it under different curvature radii.Quantitative statistics of drivers’ subjective evaluations were carried out in the two types of experiments,which verified the effectiveness of the proposed method in improving the driver’s subjective driving experience.