Hierarchical Control Strategy To Study Fuel Cell Hybrid Electric Vehicle Platoon For Complex Operating Conditions

The development of today’s automotive industry mainly revolves around safety,comfort,energy saving and environmental protection,and intelligent,networked and hybridization are currently popular research directions.In this paper,the research object is to improve the safety and economy of fuel cell hybrid electric vehicles(FCHEV)platoon in an intelligent networked environment,and the research objective is to improve the safety and economy of the platoon under complex road conditions.This paper proposes a hierarchical optimization control strategy for platoon with IPSO,MPC and reinforcement learning as the main research methods,and verifies the feasibility of the strategy through simulation test experiments.This paper is based on the National Foundation of China "Research on robust control of dynamic lane change multimode switching of intelligent heavy duty Semi-Trailer for complex traffic environment"(52262053)and "Research on stability control strategy of heavy duty semi-trailer in complex conditions of mountain highway considering dynamic changes of road parameters"(202101AT070108)of the Yunnan Provincial Science and Technology Program in 2021,the specific research includes:1)By analyzing the hybrid power system configuration of FCHEV and combining with the main usage scenarios of FCHEV,the hybrid power system is determined as fuel cell-power cell configuration.According to the system configuration,the key components of the powertrain,such as drive motor,power cell and fuel cell,are analyzed and selected to match the model,so as to build the model of the whole vehicle.Finally,the main evaluation indexes of the powertrain used for the analysis data are determined according to the research purpose.2)The upper longitudinal controller of FCHEV based on MPC is built.Firstly,the kinematics model of FCHEV is established by utilizing linear feedback theory,upper control objectives and constraints.Then,by analyzing the vehicle spacing strategy,in order to give consideration to traffic capacity and safety,variable headway is selected as the core algorithm of upper control,and the state space model of V2 V system is established through control objectives and constraints.The basic principle of MPC,state space model derivation,penalty function and Lyapunov stability are introduced in detail.Finally,the upper control strategy based on MPC is established by Matlab/Simulink,and the simulation verification is carried out by using three special working conditions: steady acceleration,deceleration,sudden acceleration,deceleration and emergency insertion of the preceding vehicle into the platoon.The simulation results show the effectiveness of the proposed strategy.3)The lower-level energy management controller based on Q-learning(QL)is built.First,the relationship between reinforcement learning and Markov decision making is introduced,and then the principle and basic components of reinforcement learning are outlined in detail.By comparing and analyzing different algorithms of reinforcement learning,the QL,which is easier to converge,is selected to build the lower layer control strategy to ensure that the lower layer controller can guarantee reasonable power allocation at any moment based on the upper layer feedback.Finally,the lower layer controller is built by Matlab/Simulink,and the validation conditions are set as HWFET conditions and complex slope conditions.The simulation results show that the QL-based strategy effectively improves the energy economy4)To improve the adaptability of the platoon vehicles in the complex dynamic environment under the intelligent networked environment and reduce the energy consumption of the vehicles during operation,a hierarchical control strategy that integrates the driving safety,tracking performance and fuel economy is designed.First,the upper layer controller solves the desired optimal vehicle speed offline by IPSO algorithm on the basis of ensuring the safe driving of the platoon,and uses it as the ideal reference input of MPC to ensure the optimality of the upper layer speed planning.Then,the lower layer control system implements energy management for the hybrid vehicle based on the optimal speed and demand power based on the QL strategy.Finally,simulation verification shows that the proposed strategy can reduce energy consumption,reduce cost,slow down battery degradation and extend battery life while ensuring good queueing tracking performance and safety.