| With the rapid development of control,sensing,communication and electronic technology,the automobile industry has gradually moved from pure mechanical system more than a hundred years ago to information physics system.A large number of advanced sensing and control algorithms have been applied to vehicles.The automobile industry has entered the digital age.Electronic control steering technology has become the main development direction of automobile steering system.At present,the automobile is popularized rapidly all over the world,and the Electric Power Steering system(EPS)is widely used in the international automobile.With the development of science and technology and the continuous improvement of material living standards,people put forward higher requirements for the driving quality of automobile steering,which not only requires the car to be easy to drive,including when there is external interference,the car has good robustness,does not need the driver to correct the steering frequently,the driver can control the different models according to the existing driving experience,but also requires the driving feeling of the car to feel good and to provide comfortable steering power feedback.After a period of use or after driving on bumpy roads,the steering performance of the car will not decline,and the driving feeling will not change.Driving quality becomes one of the core competitiveness of international automobile products.In order to ensure that the automobile has good driving quality,on the one hand,the automobile manufacturers rely on the engineering calibration method,usually through the subjective evaluation of the vehicle filed test to modify the EPS assistance characteristics in the last stage of automobile product development,so as to adjust the steering feeling of the vehicle.However,the classical EPS directly calculates the steering assist torque through the nonlinear two-dimensional field of the steering wheel torque and vehicle speed.It combines the driving intention of human beings and the dynamic characteristics of the vehicle,increases the difficulty of steering feeling calibration,the calibration workload is large,and the cost is high.On the other hand,automobile manufacturers have improved the manufacturing,assembly accuracy and wheel positioning accuracy of automobile chassis system,and auto parts enterprises have also continuously improved the machining accuracy requirements of parts.The steering machine adopts precision machining to ensure the production consistency of the steering system.However,the classical EPS realizes the assist through open-loop control,and after a period of use or collision,the performance of the functional components of the vehicle will change,which will lead to the change of steering characteristics and it is difficult to ensure the consistency of steering performance.Therefore,the use of low-cost development and manufacturing platform to make cars with high level of driving quality has become one of the development vision and technical problems of international cars.In this paper,a decoupling method of EPS assistance characteristics is proposed to solve the problem that the classical EPS directly calculates the assist torque through the nonlinear field of steering wheel torque and vehicle speed,which couples the driving intention of human and the dynamic characteristics of vehicle,and increases the difficulty of calibrating steering feeling.The proposed decomposition method provides a theoretical basis for the design of EPS assistance characteristic and reduces the calibration difficulty of steering feeling.In view of the fact that the classical EPS can not solve the problem of vehicle steering performance degradation by using open loop control method,based on the existing steering assistant mechanism and vehicle sensing,a EPS control method based on steering angle closed loop is proposed to improve the driving quality and steering performance consistency of automobile steering,and a EPS control strategy based on pavement type classification is proposed to improve the steering road sense and stability on different types of pavement.The main research contents of this paper are as follows:Firstly,the vehicle steering dynamics model and the classical EPS control model with complete functions are established.Through the comparative analysis of simulation and real vehicle filed test,the accuracy of the model is verified,and the virtual test is provided for the test and calibration of vehicle dynamics characteristics,the function development of electronic control system and the adjustment of automobile steering feeling.Through virtual test instead of real vehicle filed test,the adjustment cycle is greatly shortened,the matching difficulty of vehicle type is reduced,and the development cost is saved.Secondly,study on dynamic mechanism and calculation method of steady state characteristics of steering assist torque.In order to solve the problem that the classical EPS calculates the assist torque directly by the nonlinear field of steering wheel torque and vehicle speed,coupling the driving intention of human and the dynamic characteristics of vehicle,the difficulty of steering sensory calibration is increased.Firstly,the perceptual mechanism of human body to automobile motion is analyzed from the physiological point of view,and the steering operation behavior of driver is analyzed from the point of view of control,and then the lateral acceleration and steering angle are taken as intermediate variables.The EPS assistance characteristics are divided into three functional modules: driving style,inverse characteristics of vehicle handling dynamics and inverse characteristics of steering system dynamics.Based on the model,the steady-state inverse characteristics of vehicle handling dynamics and steering system dynamics are calibrated.The existing actual vehicle driving style is identified by differential evolution algorithm,which provides a theoretical basis for the design of EPS assistance characteristics and reduces the difficulty of steering feeling design and adjustment.Thirdly,the EPS control method based on steering angle closed loop is studied.In view of the problem that the classical EPS can not solve the problem of automobile steering performance degradation by using open-loop control method,based on the internal model principle,a closed-loop control method of EPS steering angle with direct feed and forward channel inverse system correction is proposed.The transfer function of the equivalent controlled system is identified by using Gauss-Newton iterative algorithm,and the multi-objective optimization method of internal model correction parameters is explored.The robust design of controller parameters is carried out by the nonsmooth optimization algorithm.The proposed control method unifies steering assist and return control without additional dynamic compensation,reduces the design,manufacturing and assembly accuracy requirements of steering gear,ensures the consistency of steering performance,and makes it possible for cars built on low-cost development platforms to have a high level of driving quality.Fourth,the steering road feeling control on different types of pavement.An EPS control strategy based on pavement type classification is proposed.Based on the typical pavement experience data,the pavement type is identified,and the control target is modified according to different pavement types,which improves the steering road sense and steering stability on different types of pavement.The experimental results show that the algorithm has good working condition adaptability.Finally,the automobile driving style is designed through the ideal steering feeling,and the effectiveness and structural superiority of the EPS control method in the consistent control of automobile steering performance are verified by experiments in the driving simulator environment,which can effectively improve the driving quality of the vehicle,have good adaptability to the working conditions,and can effectively reduce the design and manufacturing difficulty of the steering gear,which make it possible for vehicles built on low-cost development platforms to have a high level of driving quality. |