Research On Unmanned Truck Fleet Scheduling Method Based On Port Cloud Background System

With the rapid development of science and technology in recent years,automatic driving has gradually replaced the work of drivers,which can solve the problem of traffic safety greatly.Among the many application scenarios of autonomous driving,commercial vehicles in limited scenarios can be commercialized earlier than passenger vehicles in the open road.The direction of autonomous driving in the park has become a research hotspot in this time.In various park scenarios,the container terminal in the port has complete infrastructures and standardized roads,which are suitable for promoting the development of autonomous driving technology.In addition,the port is currently facing the problem of a shortage of truck drivers.The development of unmanned truck fleets and intelligent transformation of the port cannot be delayed.With the application of 5G communication technology,this paper studies the scheduling control method of the unmanned fleet in the cloud port vehicles manage system based on the Internet of Vehicles,and It is verified on the port trial operation and simulation platform.The main works as follows:Firstly,aiming at the guidance method of unmanned trucks,because of the complexity of trailer local path planning and vehicle control,a method of transmitting the virtual path composed of the key points through the Internet of Vehicles is proposed,which replaces the traditional scheme of laying navigation magnetic nails in the field.Byanalyzing the historical trajectory of the human truck driver,the key points of the virtual path are generated in the fitting curve described by the k-order polynomial.Secondly,for the global path planning problem of unmanned trucks,the road network data of the port high-precision map is used to abstract the data structure of a directed graph with lanes as the smallest unit node,and the Dijkstra algorithm is used to generate the global path sample library under the static road network.The vehicle attitude in Frenet coordinate system is used to locate the lane node in the road network.Furthermore,a dynamic road network model which can reflect the state of traffic flow at the lane-level is proposed,which makes the cost of global path planning not only consider the pathmileage.Thirdly,it proposes the task management method and fleet scheduling method for unmanned trucks in the port background system.Using the finite state machine model,the task flow of unmanned trucks is automatically jumped and the fault recovery logic is realized.Using the finite state machine(FSM)model,the logic for unmanned trucks about the task flow automatically jumped and the fault recovery is realized.Based on the rule engine,the multivehicle collision detection and avoidance functions in intersection are realized combined with the dynamic window method;the lane change trigger in the yard area is realized combined with the safe lane-change model.Finally,the function is verified based on the port’s actual unmanned truck fleet and simulation platform.The correctness of the global path planning in the process of typical operation is verified.And through the data of the trial operation stage,Analyzing and proving the effectiveness of the port cloud background system for truck fleet scheduling.