Research On Motion Control Strategy For Extreme Maneuvers Of Self-driving Vehicle

Many traffic accidents are mainly caused by the driver’s failure to control the vehicle under the friction limit.Most of the existing motion control technologies for self-driving vehicles design control algorithms to avoid the vehicle reaching the friction limit.However,by observing the driver’s control of the vehicle,it is found that the driver often races under the limit of vehicle friction without losing control of the vehicle.Many fatal accidents can be avoided if autonomous vehicles or driver assistance systems have similar capabilities to racing drivers.In order to ensure safety in a wide range of situations,exploring the motion control of self-driving vehicles under extreme conditions is helpful to improve the driving safety of self-driving vehicles.Therefore,based on the national key R&D plan "new energy vehicle" special project,this paper studies the stable tracking condition and drift condition of the selfdriving vehicle under the extreme condition of tire friction limit.First of all,it is very important to establish a tire dynamics model with appropriate accuracy.The tire modeling should consider the model accuracy comprehensively,but at the same time,considering the vehicle running under the friction limit,there is a high demand for real-time control,and the tire model also needs to consider the computational complexity.A brush tire model considering the tire cornering and longitudinal slip was established.A three-degree-of-freedom vehicle model was established.Secondly,on the basis of modeling,aiming at the problem that it is difficult to guarantee the tracking control accuracy and stability of self-driving vehicle under extreme conditions,a comprehensive coordinated control method of longitudinal and transverse stability is proposed.The speed of the self-driving vehicle under the friction limit is planned,and the speed following under the limit speed is realized by longitudinal feedforward and state feedback controller.The preview feedforward and artificial potential field feedback are combined to design the lateral path tracking controller.Then,a stability control strategy is designed to optimize the driving torque of longitudinal control by the deviation between the expected and actual yaw rate.The Simulink / Carsim co simulation results show that the proposed method can improve the transient response of the self-driving vehicle under the limit conditions,restrain the overshoot at the road curvature mutation,reduce the steady-state error in path following,and improve the trajectory tracking accuracy of the self-driving vehicle and the lateral stability in the process of curve motion.Thirdly,aiming at the drift condition,the dynamics of the drift condition of the friction limit of the self-driving vehicle is deeply analyzed,and the steady-state equilibrium points of the three degree of freedom vehicle dynamics model under the given front wheel angle and longitudinal speed are calculated,including the typical steady-state points and the drift steady-state equilibrium points.Secondly,the dynamic information around the equilibrium point is analyzed through the phase diagram.It can be seen that the typical characteristics of vehicle drift are reverse front wheel angle,large mass center sideslip angle and obvious rear wheel slip,so the steady-state equilibrium point of drift corresponds to the saddle point in the phase diagram.Because the rear tire is always saturated at the drift equilibrium point,the longitudinal force and lateral force of the tire are coupled,so the rear wheel drive can control the lateral dynamics at the equilibrium point.Then,based on the model predictive control,combined with the variable step discrete vehicle model and considering the rear wheel speed,a variable step MPC drift controller based on wheel speed dynamics is designed to make the vehicle steady-state circle drift.Based on Matlab / Carsim co simulation,it has been shown that the variable step size discretization can ensure the error accuracy and shorten the calculation optimization time,and the wheel speed dynamics control can reduce the control jitter,thus verifying the effectiveness of the proposed control algorithm.Finally,for further verification the effectiveness of the designed extreme maneuvers motion control algorithm in the real controller,the HiL experimental platform of self-driving vehicle is built.The experimental platform consists of the following three modules: the host computer system,D2P rapid prototyping development platform and NI real-time simulator.Firstly,the vehicle dynamics model is built and written into NI real-time simulator through NI veristand software,and then the control algorithm is written into the real D2P controller to complete the experiments of two motion control algorithms under limit conditions.The results show that the designed control algorithm can meet the requirements of robustness and real-time in the real controller.