The Research Of Robot Twist Lock Automatic Assembly And Grasp Planning

Modern logistics industry develops rapidly.The automation of container load is improved with the improvement of Container load rate.However,the assembly of twist lock is also need manual labor.In order to improve the automation of container load,a solution of installing robots to replace manual labor to finish the assembly and grasp of twist lock is proposed,which is based on the demand of container load logistics automation.This article uses ABB robot as research object.But the function of ABB off-line programming software on twist lock assembly and grasp is simple,particularly on grasp and assembly.On one hand,off-line grogramming software can not plan path flexibly for complex operating environment,it is hard to change the robot path.On the other hand,when assembly twist lock,it can only use fixed model,there is no way to assembly intelligently on working field.According to above problems,a virtual simulation lab is established by the research of the secondary development of RobotStudio off-line programming software and machine learning algorithm.In the lab,grasp and anssembly can be simulated.With the help of ABB robot platform,the correctness of container twist lock grasp and assembly solution can be verified.The secondary development of RobotStudio focus on target points setting and path planning.Using quaternion to simplify the input of target points by analyzing the structure of target points data.Arranging target points data in order by Queue,so that the path can be created with the motion of robot.And target points are labled to control the robot to finish actions like grasp,carry,assembly and so on.The research of machine learning algorithm is used to the assemble process of twist lock and container corner.Combine six-dimensional force sensor,a learning model is established by principal component analysis(PCA),clustering analysis and hidden markov model(HMM).The model analyses,disposals,learns sensor feedback data to judge the position relationship of twist lock and comer.And then controls robot to finish intellgent assembly in cooperation with secondary development programm.In order to verify the applicability of secondary development program and machine learning algorithm,setting up a virtual simulation lab and leading in twist lock gripper,corner and other equipments to simulating grasp,carry and assembly twist lock.Finally,actual control experiment of robot motion path on ABB IRB120 robot platform is carried out to demonstrate the effectiveness of the proposed scheme.