| Mobile manipulator,with their unique flexibility and flexibility,has been widely used in 3C electronics,semiconductor,medical,daily chemicals,agriculture,machining,and other fields,but there are few application cases in the field of fabric.In the field of fabrics,fabric samples are typically displayed hanging in multiple fabric cabinets.At present,there are the following problems in the storage and sorting of fabric samples:The warehouse layout is unreasonable,and the sorting of fabric samples is manual,with high costs and low efficiency.It is impossible to achieve refined management and intelligent sorting.With the continuous improvement of labor costs and intelligent level,intelligent sorting of fabric samples has become imperative,and innovative solutions based on information warehousing and intelligent sorting need to be sought.Due to the fact that fabric samples are stored in fabric cabinets at different locations,and the arrangement of fabric samples in the same fabric cabinet is relatively dense,there are certain requirements for intelligent sorting of fabric samples.The combination of "hands" and "feet" of the mobile manipulator provides mobility and sorting functions.With machine vision,it can achieve precise positioning and refined operations,and can meet the functional requirements of intelligent sorting of fabric samples in different regions.Therefore,mobile manipulator has broad application prospects in the field of intelligent sorting of fabric samples.This paper takes the fabric sample sorting as the application scene,and designs a mobile manipulator to realize the intelligent sorting of fabric samples.The process flow of fabric sample sorting by mobile manipulator is as follows:when the mobile manipulator moves to the designated fabric sample display cabinet,the cooperative robot is controlled to move near the fabric sample.The identification and pose estimation of the fabric sample are completed by the depth camera.After reading the pose data,the cooperative robot adjusts the pose of the electric gripper to accurately complete the sorting of the fabric sample and place it on the vehicle fabric rack.Based on the above sorting process,the research focus of this paper includes four aspects:mobile manipulator development plan and ontology design,mobile manipulator kinematics analysis,fabric sample sorting path planning based on MoveIt!,fabric sample pose estimation and sorting experiments.The specific research content is as follows:First of all,carry out mobile manipulator development plan and ontology design.Based on the process flow of fabric sample sorting and the application goal of functional requirements,this paper has completed the project design and ontology development of a mobile manipulator.The whole constitution of the mobile manipulator is divided into three units:mobile platform unit,robot sorting unit and vehicle-mounted fabric rack unit,which respectively realize the functions of autonomous positioning and navigation,fabric sample identification and sorting,and storage of sorted fabric samples.The controllers are selected and the construction of the control system is built according to the division of the control module of the mobile manipulator.Secondly,the kinematics analysis of the mobile manipulator is carried on to offer theoretical guidance for motion control of the mobile manipulator.In this paper,the method of decentralized motion control is used in the mobile manipulator,so the mathematical models of the cooperative robot and the mobile platform are established respectively for forward and inverse kinematics analysis,and the kinematics analysis of the cooperative robot is verified by Matlab.Thirdly,through MoveIt!Jointly simulates the sorting process with Gazebo to check the stability of the mobile manipulator during sorting,and finally MoveIt!is used to realize the motion control of the real cooperative robot.The simulation model of cooperative robot is established under ROS operating system,and the simulation model is adopted by MoveIt!.The motion path of the cooperative robot is planned and demonstrated by the 3D visualization tool Rviz.In order to better simulate the grasping process in reality and verify the stability of mobile manipulator grasping,this paper adopts MoveIt!co-simulation with Gazebo is carried out to verify the running effect of the model,and finally this paper implement motion control for real cooperative robot by MoveIt!.Finally,a method combining two-dimensional grasping pose and depth values is proposed for fabric sample pose detection,and intelligent sorting experiments of fabric samples are completed using a mobile manipulator.As the fabric samples in the scene are thin and densely arranged,and the occlusion is serious,this paper obtains the coordinates(x,y)of pixel grasping points and grasping angle θ through a two-dimensional grasping detection method based on directed arrows,and then gains the three-dimensional coordinates(x,y,z)of the point in the camera coordinate system through depth value and the pixel grab point(x,y).The above data are integrated and converted into pose data under the coordinate system of the cooperative robot base,and the intelligent sorting of fabric samples is realized by the mobile manipulator after reading pose data to verify the feasibility of the scheme in this paper. |
