Research On Human-Machine Shared Lateral Control Strategy Based On Consistency Of Risk Assessment For Intelligent Vehicle

With the rapid development of autonomous driving technology,autonomous driving systems have gradually shown their advantages in precise perception and control capabilities.However,the application of global autonomous driving still faces many challenges at this stage,and the driver still plays an important role in the car.The human-machine shared driving in which the driver and the automatic driving system drive together is a necessary stage for the development of intelligent vehicles.How to design a reasonable strategy to realize humanmachine shared collaborative control,use the respective advantages of the driver and the automatic driving system to improve the intelligent vehicle ’s ability to perceive and deal with risks in the traffic environment,and realize the safe and stable driving of the vehicle in the complex traffic environment.It is a key problem that human-machine shared driving needs to solve.This paper proposes a human-machine shared lateral control strategy based on the consistency of risk assessment.The consistency of the driver’s and the automatic driving system’s perception of risk in the traffic environment is used as the main basis for the allocation of driving rights in human-machine shared driving.Considering the difference of humanmachine driving operation,the driving weight is assigned based on fuzzy rules.Considering the dynamically changing driving rights and driver input,the lateral controller based on model predictive control is designed to realize the shared lateral control between the driver and the automatic driving system,ensure the safe driving of the vehicle,and optimize the driving trajectory of the vehicle.This paper mainly focuses on the following four aspects:(1)Driving safety field modeling and lane change trajectory planningFor lane changing scenarios,the risks generated by various traffic elements in the traffic environment are described in the form of driving safety fields,including the potential energy field generated by stationary objects and the kinetic energy field generated by moving objects.The decision-making process of vehicle lane-changing is analyzed,and it is divided into two stages: lane-changing intention generation and lane-changing feasibility judgment.According to the initial state and the end state,a quintic polynomial is used to plan the lane-changing trajectory in the frenet coordinate system,and the optimal trajectory is selected as the target trajectory tracked by the controller of the vehicle automatic driving system during the lanechanging process.(2)Research on consistency of human-machine risk assessmentThe driver’s perception is analyzed,and the driver’s eye movement information collected by the eye tracker is used as the basis to judge the effective area of the driver’s risk perception.The driving safety field of the perception result of the automatic driving system is projected onto the driver’s field of vision plane through coordinate transformation,and a twodimensional driving safety field image with field strength information matching the scene image collected by the eye tracker is obtained.Then,through the matching calculation between the effective area of the driver’s risk perception and the area where the risk perception result of the vehicle automatic driving system is located,the consistency of the human-machine risk assessment is obtained.(3)Human-machine shared lateral control strategyTaking the consistency of human-machine risk assessment as the main basis for the allocation of driving rights,and considering the differences in human-machine driving operations(referred to as the front wheel steering angle in lateral control conditions),the design takes these two indicators as the input and the driver’s driving weight as the output.The fuzzy rule controller realizes the online allocation of driving rights.A model predictive control-based lateral controller is designed according to dynamically assigned driving weights and driver inputs,and the tracking target is the trajectory planned by the autonomous driving system.The control amount is the front wheel steering angle of the automatic driving system,and a steering angle constraint that changes with the driving right is set.Based on the optimization solution of the objective function and the constraint,the control result and the driver’s operation are weighted according to the driving weight to achieve control fusion.(4)Driver-in-the-loop experiment verificationThe driver-in-the-loop experiment is carried out on the driving simulator platform to verify the lateral control strategy designed in this paper.Build a driving simulator platform,simulate the real driving environment,collect driver information,and realize the information interaction of traffic elements such as "people-vehicle-road".Analyze the verification requirements,and set the corresponding experiment scenarios according to the perception results of the driver and the automatic driving system and the different target trajectories.The driver-in-the-loop experiment is carried out,and the experiment results are analyzed to verify that the control strategy in this paper realizes the human-machine shared control to ensure the safe driving of the vehicle.