Study On Adaptive Steering For Intelligent Four-wheel Independently Driven In-wheel Motors Electric Vehicle

Nowadays, the developments of vehicles are tending to electric and intelligent direction. The four-wheel independent drive in-wheel motor electric vehicle is the important development direction, and its control has become a research hotspot. But most research work is for the vehicle chassis system at present, and it does not take into account the drivers features. The paper studied on four-wheel independent drive in-wheels electric vehicle control with consider the drivers features based on the National Youth Natural Science Foundation of China(Project No.51305190) which is the adaptive steering research for intelligent four-wheel independent drive in-wheel motor electric vehicle.Firstly, the classification method for drivers steering characteristics is studied. The multiple rectangular turning experiment conditions are designed for driving simulator experiment, and some experienced drivers are chose. Some data are collected which can reflect the driver steering characteristics. The drivers steering characteristics are eventually divided into cautious type, general type and radical type three categories by using fuzzy C-means clustering method. Secondly, based on the reasonable classification of drivers steering characteristics, the driver steering characteristics identification model was established by using BP neural network. The characteristic values in corners are selected as the input of identification model, the driver type as the output of identification model. The network structure is designed, and the identification model is trained. In the end, the model identification accuracy is verified. On this basis, three kinds of experimental data are collected through the driving simulator experiment. The reference models of different driver steering characteristics are established by using RBF neural network method. The network structure is designed, and the model is trained. Those reference models are verified offline. Then, based on adaptive fuzzy control theory, the yawing moment parameter self-regulation fuzzy controller was designed. The additional yawing moment was distributed with four-wheel drive rules allocation methods. The last, based on the driving simulator, the offline training characteristics identification model and the reference model are embedded in the vehicle driving simulator model system. The different types drivers are selected to experiment on the driving simulator. The identification accuracy of drivers steering characteristics and the four-wheel drive control effect after matching reference model were verified online.The verification results showing that the research method can realize the accurate identification, automatic matching the characteristics of the driver steering performance reference model, and adaptive steering control of electric vehicle was realized by four-wheel drive torque control.