Development of Integrated Arc-Welding System with Visual Faculty Based on Intelligent Utility of Production Information

本研究开发的系统由焊接条件数据库和视觉系统、焊接机器人系统、装配机器人系统三个子系统组成。研究结果总结如下：1.研究成果。用IF-THEN式语句表示的模糊规则有效地表达了焊接过程的知识。该模糊规则是在模糊空间中单独定义的，通过将实验焊接数据以图表的形式进行划分。版画作品的装配件之间的形状和相互关系可以用PLLOG谓词来表达。说明该表达式适用于装配序列生成等知识处理。提出了一种新型的火炬位置确定路径搜索法。该方法在机器人手臂的焊接位置x和姿态角Rheta表示的x-Rheta平面上搜索路径。阐明了在给定焊接姿势条件下的焊接工艺顺序很容易生成。阐明了利用机器人起点到末端执行器之间的距离标准化的手势地图，可以很容易地确定装配机器人的姿态和路径。利用新研制的狭缝激光扫描莫尔图像和灰度图像构建了视觉系统。该视觉系统对板形数据的获取范围更广，获取的信息可靠性更高。物体形状的轮廓可以用每个表面形状来表示，并且表面形状随着视觉方向的变化很小。本研究利用基于曲面形状的数据结构，实现了高速的形状匹配。

The system developed in this research is consisted with the database of the welding condition and three sub-systems ; vision system, welding robot system and assembly robot system. The results are summarized as followings.1. The knowledge on the welding process is expressed efficiently by the fuzzy rule expressed by "if-then" type sentence. This fuzzy rule is defined individually in the fuzzy space where the experimental welding data is divided by putting into the form of a diagram.2. The shape and co-relation between assembly parts of the works constructed by plate works are able to express by the plolog predicate. It was clarified that this expression is fit for the knowledge processing such as generation of assembly sequence.3. New type path-searching method for decision of the torch position was proposed. In this method, the path is searched on x-rheta plane expressed by the welding position x and posture angle rheta of the robot arm. It was clarified that the sequence of the welding procedure under the condition of the appointed welding posture is generated easily.4. It was clarified that the posture and path of the assembly robot is decided easily by using the map of the hand posture standardized by the distance between the robot origine and the end-effector.5. The vision system using a new developed slit laser scanning moire image and a gray scale image at the same time was constructed. This vision system has wider obtaining range of the shape data and higher reliability of the obtained information.6. Outline of the object shape can be expressed by each surface shape, and the surface shape is scarecelly changed according to the visual direction. In this research, high-speed shape matching was realized by utilizing the data structure based on the surface shape.