用光学阵列传感器的机器人物体分类系统

本文介绍用光学阵列传感器的机器人物体分类系统。传感器直接安装在机器人的两个手指上。被抓物体的阴影通过光导纤维传到安放在“安全区”的光敏元件上。计算机识别物体的轮廓后命令机器人抓握物体,并把它运送到指定的地点从而达到物体分类的目的。

A low-cost Optical Sensory sorting system is described in this paper.The sensoris mounted directly on the robot gripper fingers.It uses a light source and a bundleof optical fibers on opposing robot fingers.The optical fibers carry the shadow of agripped object,as an eight-by-eight pixel array,to detection,multiplexing,discri-mination,and computer interface electronics mounted on the robot base.The systemuses a microcomputer for several data processing and pattern recognition functions.Discussion in this paper includes the design and analysis of the sensor and its opti-mal array,the hardware,and the parts recognition and control system.System per-formance in a task which requires acquisition,identification,and bin-sorting of avariety of electronic and mechanical parts is also covered.