Acceleration Process Optimal Control Of An Electric Vehicle Based On Driving Styles

With the rapid development of electric vehicle technology,four-wheel-drive electric vehicle has drawn increasing attention of many researchers from various field,due to its more flexible configuration and its independently controllable wheel torque.However the problem is that its jerk always performances large during the launch and acceleration process.Currently,most studies concerning this issue have ignored the will of drivers,which is actually important for vehicle performance.Oriented by the concept of driving style,this paper researches the multi-objectives optimal control of torque coordination distribution as follows,considering vehicles' stability,comfort,dynamic performance and economy performance.(1)First of all,based on three classified driving styles which are obtained by the proposed improved K means algorithm,combined all parameter related to each objective,this paper determines the principal optimization objective,secondary optimization objectives and corresponding parameters of each driving style.Considering the characteristics of four-wheel-drive electric vehicle,the driving style related parameters is then transferred to get the value of parameters corresponding to all objectives.Consequently,the control system obtains the reference curves of the plant output by polynomial fitting the experimental data.(2)Afterwards,based on vehicle's dynamic model,this paper constructs two vehicle models for the proposed optimal controller as internal model and plant model respectively,including vehicle body model,motor model and tire model et.al.The accuracy of plant model is test by the comparison with CarSim model.(3)In addition,aiming at improving the global performance during acceleration and launch,we proposed a control strategy to trade off the conflicting objectives based on the acceleration curves obtained by polynomial fitting and the constructed internal model.This strategy realizes driving-style-oriented multi-objectives optimal control by adopting different output reference curves and distinct cost functions.Leading by this strategy,the proposed optimal controller calculates four desired torque vectors of motors under each driving style to facilitate a better acceleration.(4)Finally yet importantly,four indexes are determined according to the property of each objective so that the performance of vehicle can be evaluated.Simulation scenario selects dry asphalt road,snow road and several spilt friction roads to test the feasibility and robustness of the proposed controller.Simulation outcomes of three driving styles are compared at last.