Research On Self-learing Strategy And Control Method Of Eletric Vehicle Hub Driving Robot

The dynamometer test is a very important indoor bench test in the development process of automobile products.Through the dynamometer test,not only the dynamic performance of electric vehicles can be tested,but also the energy consumption index of multiple working conditions can be measured.Performances such as power,economy and comfort provide reference.The hub driving robot for electric vehicles can replace the human driver to perform the dynamometer test,which effectively improves the test accuracy and saves test cost.In the paper,based on the basic requirements of speed tracking control,which is the basic requirement of the hub test,the self-learning strategy and control method of the electric vehicle hub driving robot are studied.First,based on the electric vehicle longitudinal control system model,the electric vehicle simulation software was developed,combined with the Beckhoff PLC controller,and the electric vehicle driving robot semi-physical simulation platform was built.Among them,the Beckhoff PLC controller runs the speed control algorithm,and the simulation software runs the mathematical model of the electric vehicle longitudinal control system.Through the interaction of PLC and simulation software,the real hub driving robot is simulated to control the electric vehicle for dynamometer test.Based on the semi-physical simulation platform,the preliminary available parameters and PLC program of the controller can be obtained before the actual test,reducing the time for debugging on the test site.Semi-physical simulation experiments are carried out on various speed tracking strategies based on PI control combined with inverse control strategy models,which preliminarily verify the feasibility in engineering.Secondly,in view of the characteristics of the driving robot using the accelerator pedal to achieve speed tracking,the inverse control strategy model is introduced on the basis of the traditional longitudinal dynamic model based on the motor torque to achieve speed control.The inverse control strategy model realizes the mapping of the speed tracking control quantity from the motor torque to the accelerator pedal opening.Among them,the motor torque control quantity is obtained by designing the model predictive controller,and before designing the model predictive controller,it is necessary to identify the model parameters of the electric vehicle longitudinal control system from the collected real vehicle data,and input a set of real vehicle accelerator pedal opening The longitudinal control system model compares the output vehicle speed of the model with the actual vehicle speed to verify the accuracy of the model establishment.Simulation experiments are carried out on the model predictive control and the inverse control strategy model to verify the effectiveness of the method.Then a set of controller parameter self-learning method is studied,which realizes the complete process from collecting test data to optimizing the controller parameters.Among them,the electric vehicle longitudinal control system model parameters are identified based on the collected real vehicle data.Based on the system control model,it is proposed to use the feedback linearization method to accurately linearize the nonlinear longitudinal dynamics model,and then based on the classical control theory.Design the initial values of the controller parameters,and then apply the simulated annealing algorithm to optimize the controller parameters.Taking PI controller as an example,this paper conducts Matlab numerical simulation experiments on feedback linearization and simulated annealing algorithm to optimize PI parameters,which preliminarily verifies the theoretical rationality of the proposed method.Finally,the speed tracking scheme based on PI control,with strategies such as preview,acceleration pedal stability maintenance,and hub braking,was tested on two typical automobile test conditions of NEDC and Chinese automobile driving conditions.The results are It shows that the hub driving robot for electric vehicles can replace human beings to complete the dynamometer test with high efficiency and high precision.At the end of the paper,the research content of the subject is summarized and the future work direction of the subject is prospected.