Research On Longitudinal Cooperative Control Strategy Of Truck Platoons Under Highways Working Condition

With the development of the economy,China’s motor vehicle ownership continues to increase,highway truck transportation as an important part of transportation,greatly affect traffic congestion,traffic accidents,energy crisis and a series of problems,and truck platooning can better solve these problems.At present,the research on truck platoons has many problems such as large modeling errors and uncertainties in vehicle dynamics models,single communication topology between vehicle platoons,and poor adaptability of truck platoons to complex environments with parameter perturbation.In view of the above problems,this paper studies the stability of truck platoons under multi-type communication topology based on the closed-loop feedback truck platoons model.Firstly,the architecture of the vehicle control system with hierarchical control is established.The upper distributed controller decides the expected acceleration,and the lower controller is composed of the inverse dynamic model and the dynamic model,in which the inverse dynamic model outputs the expected throttle opening or brake wheel cylinder pressure into the dynamic model.In addition,the modeling error is corrected by closed-loop feedback control of the expected acceleration,and the vehicle dynamic hysteresis coefficientτ is introduced for heterogeneous vehicles to distinguish the dynamic response of heterogeneous vehicles.Then,a Model Predictive Control(MPC)is built in the upper controller.According to the longitudinal kinematics of the truck platoons,a discrete state equation is established which can be used by the model predictive controller.The state equation introduces the acceleration state of the vehicle in front to resist the disturbance of the system.The prediction model establishes performance optimization index function with safety,comfort and stability as control objectives.In addition,in order to enhance the stability,the reference trajectory is introduced to soften the control quantity.The feedback correction model is established by state equation and prediction equation to compensate the error and enhance the robustness of the model.For platoons stability in simple communication topology,there are some problems such as slow convergence speed,overshoot,vibration,drive and brake shock wave.Secondly,a reinforcement learning controller is established in parallel with the model predictive controller,and the performance optimization index function is established with safety,comfort and stability as control objectives.1)The longitudinal control Markov Decision Process(MDP)model of truck platoons is established;2)According to the Deep Deterministic Policy Gradient(DDPG),the following strategy of trucks in the safe distance model with fixed workshop time distance is trained,and the multi-objective reward function is designed;3)The cohort framework of reinforcement learning training and testing is designed to test the effectiveness of the trained policy model;4)For the platoons stability under the simple communication topology,the overshoot and vibration phenomena are significantly improved,but there are still problems such as driving and braking shock wave.Finally,the stability analysis of truck platoons parameter perturbation and multi-mode traffic topology is carried out.1)The stability of the heterogeneous truck platoons is analyzed according to the heterogeneous truck dynamics model;2)Establish and analyze Predecessor Following(PF),Predecessor-Leader Following(PLF),Two-Predecessor Following(TPF)platoons stability under communication topology;3)The results show that the driving and braking shock wave phenomenon does not appear in PLF and TPF communication topology,while the shock wave phenomenon does appear in PF communication topology.