摘要
根据机器人拆卸导弹发动机注药瓣模时对精确定位的要求,设计了一种基于视觉的瓣模螺栓识别定位方案。在此方案的基础上进行实验,实验结果表明:应用此方案能准确识别所有瓣模螺栓,定位误差均在1mm范围内,最大误差为0.750mm,满足机器人脱模装置的定位要求。为了保证拆卸瓣模时设备互不干涉,利用最小二乘法和Matlab对实验数据进行拟合仿真处理,求出工作转角α。应用实验数据和数据处理的结果控制机器人拆卸瓣模,检验视觉定位和数据处理的可靠性。分析视觉定位误差的主要来源,提出了适合工程应用的误差修正方法,有助于提高机器人脱模的可靠性。
According to the requirment of accurate positioning in the process of robot disassembling the injection valvemodules of missile engine, a programme is designed based on vision to identify and locate valve-module bolts. Experiments are carried out on the basis of this programme, and tests show that all valve-module bolts can be accurately recognized by applying this programme ,with the location range error of within lmm, and the maximum error of O.750mm, which meets the requirment of rabot demould device positioning. In order to ensure no device interference in the process of disassembling the valve-modules, experimental datas are simulated to get working angleby utilizing least-squares method and Matlab. Besides, for the sake of testing the reliability of visual location and data processing, experimental datas and the result of data processing are exploited to control the robot to disassemble the valve-modules. Error correction methods that are appropriate for engineering application have been proposed after analyzing the main sources of vision positioning error, which is conducive to improve the reliability of robot demould.
出处
《机械设计与制造》
北大核心
2017年第5期44-47,51,共5页
Machinery Design & Manufacture
基金
国家863计划资助项目(2014AA041603)
西安航天化学动力厂资助项目(Y2V6210601)
关键词
机器人脱模
视觉定位
最小二乘法
误差修正
Robot Demould
Vision Positioning
Least Square Method
Error Correction
