Energy Management Strategy Of Plug-in Hybrid Electric Vehicle Based On Driving Intention Under Mixed Traffic Flow

Energy management strategy(EMS)can optimize energy consumption,reduce emissions and improve the fuel economy of plug-in hybrid electric vehicle(PHEV)in different scenarios.One of the keys of optimization is the prediction of vehicle speed profile:predict speed profile of future interval,and then calculate the power demand,and optimize the power distribution of the engine and motor.Therefore,an intelligent prediction method of vehicle speed profile based on driving intention and lidar data is proposed in this paper.The predicted speed is used for equivalent consumption minimization strategy(ECMS)to achieve PHEV real-time energy optimization to meet driving intention.Firstly,the overall research method and route are given,and then the modeling method of the whole vehicle simulator and the subdivision function of each module are introduced to verify the speed tracking effect of the whole vehicle simulator.The whole vehicle simulator is mainly composed of driving condition module,driver module,energy management module,and vehicle and power system module.The construction of power components in vehicle and power system module is introduced in detail,mainly including the modeling of engine,two motors,and battery.Secondly,the driver’s longitudinal acceleration and deceleration intention are divided into five types,acceleration,rapid acceleration,uniform speed,deceleration and rapid deceleration.A driving intention recognition model based on gated recurrent unit(GRU)is developed.This model takes the vehicle speed,accelerator pedal opening and brake pedal force of the target vehicle as the input and the driving intentions as the output.The speed profile prediction method relying on vehicle communication can not be available in the traffic scene of mixed intelligent vehicles and traditional vehicles.The target vehicle can obtain the relative speed of the front vehicle in real time with the onboard lidar.The joint probability data association(JPDA)tracker and interacting multiple model method(IMM)are used to process the lidar data to obtain the relative speed without prior knowledge.Then,combining the driving intention with the relative speed of the vehicle ahead,the speed profile in the next second is predicted.Finally,a new equivalent factor(EF)adaptive law is constructed and used in ECMS together with the predicted speed.The optimization effects of EF constant,EF based on state of charge(SOC)feedback adjustment and EF based on the new adaptive law are compared.Experimental results show that the training accuracy of the GRU model is between 85%～95%,and the recognition accuracy can reach 88%,which can effectively identify the driver’s driving intention.The developed velocity profile prediction method has good velocity measurement accuracy.The maximum difference between the predicted speed and the actual speed shall not exceed 5.9km/h.ECMS can optimize the energy flow distribution according to the predicted speed and adaptively adjusted equivalent factor,effectively realize the instantaneous optimization of vehicle fuel consumption based on driving intention,and slightly improve the fuel economy.