Research On Material Selection Of Composite Robot Based On Laser SLAM Navigation

With the continuous development of robot technology and the transformation of traditional manufacturing to intelligent manufacturing,composite robot technology has gradually become a hot spot in the research field.In this paper,robot is designed for material transportation and feeding requirements in Industrial production workshop.The robot can be used in the production workshop of cigarette factory for drawing navigation,material identification,handling and loading of composite robots.In this paper,SLAM mapping,localization and navigation of composite robot and material identification and selection by 3D camera are studied.Firstly,the characteristics of the production workshop of the cigarette factory are Analyzed.On this basis,the design requirements of the robot are clarified,the robot control module is decomposed,the control system framework of the robot is designed.Then the experimental hardware platform is built,and the software platform used in the experiment is introduced.The kinematics model of a four-wheeled Mc Num wheel form composite robot is analyzed.Through the analysis of the advantages and disadvantages of the three environmental maps,the raster map is selected as the experimental map according to the actual needs.Secondly,the SLAM mapping method of composite robot is studied.The Hector SLAM algorithm based on graph optimization and the RBPF algorithm based on particle filtering are studied.In view of the shortcomings of the RBPF algorithm,an improved RBPF-SLAM algorithm is proposed.Based on odometer data and laser radar measurement data,the algorithm adjusts the weight of particles and carries out adaptive resampling.The improved RBPF algorithm has high accuracy of drawing in lab environment,and is suitable for Industrial production workshop,which is proved by the experimental results.Then,the Monte Carlo positioning algorithm is studied,and the adaptive Monte Carlo algorithm is used to perform global positioning and pose tracking of the robot in the map.At the same time,the global path planning A~* algorithm is studied,and the Floyd algorithm is added to optimize and improve it on the basis of the A~* algorithm.It is proved through simulation experiments that the improved A~* algorithm can reduce the number of path turns,traverse nodes and reduce the planning time.Meanwhile,the local path planning DWA algorithm is studied,and the real-time obstacle avoidance function of the DWA algorithm is verified in the simulation.Through experiments,it is proved that the improved A~* algorithm and DWA algorithm are combined to complete the path planning under the premise of real-time obstacle avoidance.Finally,the material selection is studied,and the type of material is identified by identifying the depth of the groove of the workpiece material and the height of the segment,and the centroid coordinates of the material are obtained to identify the position of the material.The 3D point cloud image of the target workpiece material is obtained by photographing the material with a 3D camera.The 3D point cloud data is preprocessed,and the point cloud voxel filter is used for down sampling,and the largest plane in the 3D point cloud image is segmented to complete the background removal.At the same time,the Euclidean clustering algorithm is analyzed.And the fast Euclidean clustering algorithm is proposed to improve the search accuracy.The fast Euclidean clustering algorithm is used to remove the point cloud impurities.The axis segment is divided by the regional growth algorithm,and the cylindrical point cloud and the plane point cloud in the slot are separated.In the end,the centroid coordinates of the workpiece material in the camera and the depth of the workpiece material slot and the height of the section where the groove is located are calculated.The identification of the material is completed,and the composite robot can select the material correctly.