基于概率神经网络图像识别的工业机器人控制  

Industrial robot control based on probablistic network image recognition

作　　者：黄玉钏[1] 王俞 李振平[1] HUANG Yuchuan;WANG Yu;LI Zhenping(Communication and Information Center,State Administration of Work Safety,Beijing 100013,China;Beijing SIASUN Electronic System Company Limited,Beijing 100038,China)

机构地区：[1]国家安全生产监督管理总局通信信息中心,北京110013 [2]北京新松佳和电子系统股份有限公司,北京100038

出　　处：《计算机应用》2018年第A02期63-66,71,共5页journal of Computer Applications

基　　金：安监部门典型灾种智能服务示范应用课题(2016YFC0803109)

摘　　要：为解决工业机器人基于视觉的自主轨迹控制,根据概率神经网络(PNN)分类识别和非均匀有理B样条(NURBS)轨迹规划的原理,提出基于PNN图像识别的工业机器人自主轨迹控制的方案。分析了工业机器人视觉系统模型,基于视觉系统模型的识别结果设计工业机器人自主轨迹控制器,提出利用视觉识别信息自动生成工业机器人NURBS轨迹的方法,最后用工业机器人模型设计工业机器人自主轨迹控制闭环系统。仿真结果表明基于PNN图像识别的工业机器人控制可以实现控制轨迹平稳转换,控制方案可以很好地实现工业机器人基于视觉的平稳、快速自主轨迹控制。In order to solve the autonomous trajectory control of industrial robots based on vision,based on the principle of PNN(Probablistic Neural Network)classification recognition and NURBS(Non-Uniform Rational B Spline)trajectory planning,an autonomous trajectory control scheme for industrial robots using the result of PNN image recognition was proposed.The visual system model of industrial robot was analyzed,the autonomous trajectory controller of industrial robot based on the recognition result of the visual system model was designed.The method of automatic generation of industrial robot NURBS trajectory using visual identification information was proposed.Finally,the closed loop system of autonomous trajectory control of industrial robot was designed with the industrial robot model.The simulation results show that the industrial robot control based on PNN image recognition can realize the smooth transformation of the control trajectory.The control scheme in this paper can realize stable and fast autonomous trajectory control of industrial robots based on vision.

关 键 词：概率神经网络 图像识别 工业机器人 非均匀有理B样条 轨迹控制 

分 类 号：TP249[自动化与计算机技术—检测技术与自动化装置]