AGV Road Recognition And Path Planning Based On Intelligent Warehousing

The arrival of the intelligent era urges the logistics industry to develop towards the intelligent direction.As the key of logistics,warehousing directly affects the whole logistics level and operation efficiency.The object of this paper is AGV,which is the core component of intelligent transportation system.Because of its good research value and application prospect in commercial and military fields,AGV has attracted extensive attention of a large number of researchers.In view of this situation,this paper analyzes and studies AGV vehicles from three parts: road identification technology,path planning capability and warehouse modeling and simulation.The main contents include:(1)Road identification is a functional module proposed for the AGV's position error or deviation from the road during the execution of the path.A set of road recognition scheme using lidar to detect the reflector is designed,and the layout of the reflector based on the intelligent storage topology is introduced.An AGV dynamic position recognition scheme based on mathematical geometry is designed.Under the condition of map attribute,the road correction scheme based on limiting amplitude or PID adjustment can prevent AGV from deviating from the ideal road in the process of moving.(2)For the path planning,an improved A* algorithm is proposed,which adopts the scheme of splitting and filtering scheme to increase the turning weight in the heuristic function of unknown nodes,reduce the total energy consumption in the transportation path,shorten the total length of the path and reduce the number of turns of AGV.The task splitting scheme can select as many optimal paths as possible,maximize the list of optimal solutions,and increase the probability of selecting the optimal path.Moreover,the optimal solution is selected by considering the planning results including turn consumption and concurrent search again,eliminating redundant nodes and redundant inflection points in the path.The scheme of splitting tasks through multi-threading can not only greatly accelerate the time of path planning,but also solve the local optimal problem by fully traversing feasible routes between the initial point and the target point and selecting the global reasonable routes.(3)The AGV model is constructed by the URDF function plug-in of ROS system.After the configuration of relevant attributes,the model entity is automatically generated,and the path planning algorithm function package is extended in its development mode.In the visual simulation of RVIZ,the specified algorithm is used for simulation experiment to show the optimization effect of the algorithm.The experimental results show that the improved A* algorithm is better than the traditional A* algorithm in terms of planning time,total travel and turning consumption,and more in line with the requirements of the factory environment for AGV.