包装生产线用码垛机器人智能定位方法

Intelligent Positioning Method of Palletizing Robot for Packaging Production Line

中文关键词: 码垛机器人  手眼标定  模糊神经网络  粒子群算法

英文关键词: palletizing robot  hand-eye calibration  fuzzy neural network  particle swarm optimization

中文摘要:

      目的 为提高包装生产线中码垛机器人的智能定位精度，结合图像处理和智能算法设计一种码垛机器人定位方法。方法 分析码垛机器人基本结构和工作流程。介绍一种适用范围广的快速手眼标定方法，可通过图像处理得到物料实际位置。设计一种模糊神经网络控制器，用于消除实际抓取位置和理论计算位置之间偏差。采用遗传粒子群优化算法来解决控制器参数初始值优化问题。通过实验验证智能定位方法的有效性。结果 试验结果表明，实际抓取位置和理论计算位置之间偏差可控制在0.5 mm以内，实际分拣速度最高支持180个/min。结论 所述视觉码垛控制系统可实现工作区域内物料的定位和识别，几乎不会出现漏抓、误抓等情况，满足精度要求。

英文摘要:

      The work aims to design a positioning method of palletizing robot by combining image processing and intelligent algorithm, so as to improve the intelligent positioning accuracy of palletizing robot in packaging production line. The basic structure and workflow of stacking robot were analyzed. A rapid hand-eye calibration method with a wide range of application was introduced. The actual position of material was obtained through image processing. A fuzzy neural network controller was designed to eliminate the deviation between actual grasping position and theoretical calculating position. Genetic Particle Swarm Optimization algorithm was used to solve the problem in initial value optimization of controller parameters. The effectiveness of the intelligent positioning method was verified by experiments. The experimental results showed that the deviation between the actual grasping position and the theoretical calculation position could be controlled within 0.5 mm, and the maximum actual sorting speed was 180 times/min. The visual palletizing control system can realize the positioning and identification of materials in the working area, and almost no missing and wrong grasping occurs, which meets the requirements of accuracy.