Research On Energy-saving Driving Strategy Of PHEV Based On Trajectory Planning And Improved Q Learning

As car ownership continues to rise and energy shortages and traffic congestion are getting increasingly prominent,the development of new energy vehicle technology and the promotion of energy-saving driving strategy have been recognized as important ways to reduce traffic congestion and improve energy efficiency.Plug-in hybrid electric vehicle(PHEV)is seen as a viable approach to energy shortages due to its low pollution,low energy consumption and long range.The energy management system of PHEV is the core control component of PHEV powertrain,responsible for energy distribution among multiple power sources,and its fuel saving potential is closely related to the energy management strategy(EMS).Currently,EMS based on speed planning control is a hot topic in PHEV research,which can further improve the energy utilization efficiency of PHEV by combining with energy saving driving strategy.This paper takes PHEV as the research object to maximize the energy-saving potential of PHEV through the synergistic optimization of energy efficient driving and EMS.In the upper layer of energy-saving driving,the trajectory planning of lane change overtaking and sequential quadratic programming(SQP)-based speed planning strategy in the following environment are systematically studied in the energy-saving strategy.On this basis,energy-saving driving control strategy is integrated into the PHEV energy management research.And the energy management strategy based on improved Q-learning is investigated in depth,and the PHEV energy management strategy is proposed based on energy-saving driving and improved Q-learning.The specific content of this paper includes the following four parts:In the first part,vehicle lane change trajectory tracking controller with hybrid system modelling.A vehicle kinematic model and an error analysis model based on a two-degree-of-freedom model are established,aiming at the problem of lane change and overtaking of the vehicle.And an MPC trajectory tracking controller is built based on the principle of MPC.Combined with the Autonomie simulation platform,a quasi-static simplified model of the crucial power parts of PHEV: powertrain mode,engine,motor and power battery are built using Matlab/Simulink software to provide model support for the subsequent research of layered EMS.In the second part,research on the planning and tracking control of energy-efficient driving lane change trajectories.The segmented trajectory planning and tracking control strategy is proposed for PHEV lane changing,thinking about the security and comfort of vehicle driving in dynamic traffic environments.Firstly,the complex driving scenario is simplified as the traffic scenario of vehicle lane changes overtaking process.Then,the common trajectory planning algorithms are compared,and the five-order polynomial is selected as the trajectory planning method for the lane change,overtaking and merging phases.And an MPC trajectory tracking controller is used for tracking control.Lastly,a two-lane driving road with standard lane is built in Matlab/Simulink software,and the designed trajectory planning control strategy is simulated.Simulation results show that the following vehicle can complete lane change overtaking with small tracking error.In the third part,research on energy-saving driving speed planning in the following driving environment.A speed planning strategy based on SQP algorithm is designed to combine MPC to solve the optimal following speed sequence under constraints of leading car’s speed and safe distance.Finally,based on Matlab/Simulink software,simulation tests were conducted for three standard working conditions and a real-world operating condition in Kunming.The simulation results show that under the proposed control strategy,the target vehicle can always maintain a safe following distance without collision,and can ensure that the target vehicle follows the previous vehicle accurately.In the fourth part,research on hierarchical energy management strategy based on weighted double Q learning(WDQL).To further improve fuel economy performance of PHEV,a layered EMS based on WDQL is designed.The strategy treats vehicle lane change and speed following control as the upper layer control and EMS as the lower layer control.In addition,the optimal driving speed obtained in the upper control layer is taken as the input of the lower EMS to realize an organic combination of upper-level following control and lower-level EMS control for intelligent PHEV.Simulation results demonstrate a significant improvement in fuel economy for the proposed hierarchical control strategy compared to the EMS of CD/CS.