Research On Intelligent Charging Path Planning Algorithm For Electric Vehicle

With the rapid development of automobile industry,the problem of environmental pollution and energy crisis caused by traditional fuel-engine vehicle is becoming more and more obvious.As the representative of the new energy vehicles,there still exists some problems to replace the traditional internal combustion-engine vehicles,e.g.battery duration power.The electric vehicle charging time is long,however the public charging pile resources are limited."When and where to charge" would become a problem if the cars appear when running long time outside.Try to address the "charging anxiety" problem,the thesis presents an intelligent charging path planning under the network of limited charging facilities,which is aimed to improve the charging efficiency,reduce the time in queues,and promote the development of new energy vehicles.First,the relations between the State of Charging(SOC)estimation and distance relations are analyzed based on the fundamental research.Taking the actual driving process and human as the influence factor of the SOC,a Support Vector Machine(SVM)prediction model is established.The further more accurate prediction relation model of the SOC range and traveling distance is presented based on information granulation.The path planning problem could be abstracted as a Traveling Salesman Problem(TSP)within real-time road conditions.The dynamic traffic problem could be simplified as a static path planning problem based on time discretization.Then the "Dijkstra-Simulation" path planning algorithm is put forward,which can make sure to acquire the shortest path in current period.The two-section calculations algorithm is proposed based on "Fuzzy-Accurate" to solve the queuing problem of charging pile.At last,the simulation platform is constructed and used to verify the efficiency of the algorithm.