Research On Fuel Control And Intelligent Path Planning For Connected Vehicles Based On Platooning Technology

In recent years,with the advancement of air pollution,which mainly caused by the automobile exhaust,the respiratory environment that we live in is deteriorating.The problem of energy consumption and pollution caused by automobile exhaust is emerging,which attracts increasing research interests in recent years.The research content of this thesis will foucs on route planning and fuel consumptiong controlling.This thesis first makes research on energy optimization method,which takes platooning technology to line up vehicles.Platooning technology relies on V2V(vehicle to vehicle)or V2I(vehicle to infrastructure)communication.The vehicles with the same route are gathered up into a queue,which reduces the fuel consumption by reducing the air resistance.In addition,the driving scene in this thesis is divided the into differen scenarios,i.e.,urban environment and high-speed scene.Due to the existence of intersections,a vehicle consolidation strategy that allows vehicles to join in the platoon while maintaining the stability of the fleet is also proposed.Simulation results show that the energy optimization method based on platooning technology enables the vehicle reduce fuel consumption,and the longer the fleet,the more energy is saved.However,when the number of vehicles reaches a certain level,the energy conservation ratio tends to be stable.Second,this thesis proposes a new route planning scheme based on reinforcement learning,in which Q-learning is adopted for vehicles training that aims at reducing the time required to calculate the most fuel-efficient path,and increasing the efficiency of the reaction.This thesis takes the state-return function and state-action return function as the criteria for evaluation.The study serves as a high-quality travel planning service for vehicles,which allow vehicles to choose the best driving path known at the outset and destination.The state weight of the environment combines the time and distance,by which the vehicle can compute the optimal path according to the weight calculation feedback value table.In this way,the driving efficiency can be enhanced,and the waste of network resources for the time required for path judging can be reduced.Simulation results show that the strategy proposed can select the best path required more quickly and accurately,which not only decreases the utilization of network resources,but also ensures the reliability.This thesis proposes a vehicle queuing coordinated control method,and designed a route planning protocol based on Q-learning.The queuing coordination of vehicles aims at reducing fuel consumption,and the control method drives the vehicles in groups,further trains the vehicles to find the planned route through machine learning,then forms the planned route.Finally,the effectiveness of this method is demonstrated.