Research On Intelligent Heavy-load Parking AGV Navigation System

As a necessary automatic handling and unloading means to unify and adjust the discrete logistics management system,the AGV has achieved rapid development in technology.Its application scopes overs warehousing,manufacturing,port terminal airports,libraries,dangerous sites,special industries and other fields.With the rapid development of the national economy in recent years,major cities are faced with the problems of few parking spaces and difficult parking.So,designing an intelligent overloaded parking AGV to replace the traditional parking mode or combining with an intelligent three-dimensional garage will ease the parking problems to a certain extent.Good mechanical design,motion control and precise positioning are the key technologies to realize the parking function of AGV.In this paper,we use ROS open-source robot as a platform to study the intelligent overloading parking AGV navigation system,its research is divided into the following sections.According to the design requirements and application scenarios,the overall framework of navigation is designed and the functions of each module are defined.According to the requirements of heavy duty conditions,a four-wheel independent driving and independent steering AGV is designed and the kinematics analysis was carried out to realize the omni-directional movement function.Aimed at coupling problem of four motion modules,a speed compensation model based on fuzzy PD is designed to adjust the parameters to ensure the stability of the AGV movement.The odometer model of overloaded AGV and the laser sensor model are established,based on those models and control methods,the AGV is fully moved and aligned with the carrier plate with high precision.Aiming at the problem that the path using traditional path planning algorithm is not smooth,an algorithm of using fixed radius to optimize the path is proposed,and the backstepping trajectory tracking algorithm is studied.According to the layout requirements of parking spaces,the parking electronic map is designed and the traditional A* algorithm and the improved method are compared on this map to verify the feasibility of the improved method.The trajectory tracking algorithm is simulated by matlab to verify the correctness of the method.Aiming at the characteristics of laser SLAM,which is not easy to close loop under the large-scale environment with signal and similar features,a method of combining laser triangular algorithm with graph-based SLAM is proposed to solve the problem of closed loop.Using the triangulation algorithm and the features that are pre-arranged to divide the environment into sub-regions with different ID as theclosed-loop detection signals.The Gauss-Newton iterative optimization algorithm is used to optimize the front-end data so that the obtained map can accurately describe the environmental information.Finally,the comparisons between the improved method and the traditional composition method in garage are made,and the feasibility of the improved method is verified.