Research On State Estimation And Path Following Control Of Full X-by-Wire Electric Vehicles

Nowadays,electrification and intelligentization are important development trends of automotive industry,so full X-by-wire electric vehicles come into being.Full X-by-wire electric vehicle adopts four distributed in-hub motors and four steering actuators to realize the function of four-wheel independently driving(4WID)and four-wheel independently steering(4WIS).This kind of vehicle has more controllable degrees of freedom so that it is beneficial to intelligent control of vehicle.Path following control is a significant part of automotive intelligentization.Different from front-wheel-steering vehicle,the full X-by-wire electric vehicle needs coordinated control of the four wheels to follow the desired path when steering.In this condition,more complex vehicle dynamic model needs to be established to for the design of control strategy.Therefore,the path following control of full X-by-wire electric vehicles is of great importance.During the control process of path following,the state parameters of the vehicle being obtained accurately is the premise.This paper is supported by National Natural Science Fund Project "Research on Control Mechanism and Evaluation Method of Steering-by-Wire System Based on Driver Characteristics"(Number: 51575223).The research on the state estimation and path following control of full X-by-wire electric vehicles is proposed in this paper.The path following control is based on a hierarchical integrated control method,and its control state variables are estimated by the proposed state observer.The main content can be summarized as the following aspects:(1)The design of vehicle state observer.Since that the longitudinal and lateral velocity of vehicle can not be obtained directly through conventional sensors,a nonlinear dynamic model of full X-by-wire electric vehicle is established.Then a vehicle dynamics state observer based on dual unscented particle filter(DUPF)algorithm is proposed to estimate longitudinal velocity,lateral velocity and tire force.The simulation using a high-fidelity CarSim full-vehicle model and the online test of driving simulator shows that the proposed observer has high estimation accuracy and good reliability.(2)The design of upper controller.To realize path following control,a path following control method based on preview theory is introduced,and another approach based on nonlinear model predictive control(NMPC)is designed.The controller based on nonlinear predictive model introduces longitudinal velocity,lateral velocity,yaw rate and the tracking error as state variables to improve the accuracy and stability of path following.The simulation results of three different scenarios showed that the controller based on NMPC can ensure the vehicle track the desired path more accurately and more stably.(3)The design of lower controller.Since that the wheels’ torque and steering angle of full X-by-wire electric vehicle can be controlled independently,the tires’ force under the vehicle’s coordinate are distributed based on optimal tire utilization ratio with the front,rear axle lateral force and desired longitudinal acceleration obtained by upper controller.Then with Newton iteration method,the steering angle and actuating torque can be obtained.Finally,offline simulation is carried out in two scenarios: the emergency obstacle avoidance with high speed braking operation and double lane change in low tire-road friction coefficient.The results show: the proposed path following method based on hierarchical integrated control can ensure the vehicle track the desired path and desired speed stably,so that it can realize the longitudinal and lateral coordinate control.;the proposed method of tire force distribution can ensure each wheel have a relatively high adhesion utilization margin to avoid longitudinal and lateral tire slip.Therefore,the proposed tire force distribution method can improve the anti-interference of vehicle.