Sensing and responding to the changes of geometric surfaces in flexible manufacturing and assembly

Abstract

A flexible manufacturing system is capable of dealing with changes and uncertainties occurring in a manufacturing environment. A rigid automation system adapts changes through reconfiguration or manual modification of its control programmes. In contrast, an advanced system responds to the changes autonomously or with a minimal manual intervention. One critical technology for a flexible manufacturing system is to detect or predict the changes and uncertainties and take them into account for the control of the manufacturing system. In this paper, an intelligent manufacturing system is proposed to accommodate product geometric variants autonomously by integrating a 3D vision system into manufacturing processes. The proposed system is capable of acquiring vision sensor data, detecting the geometric changes, and modifying the control programmes of the machine in response to the changes. In this innovative work, a new method for converting the raw vision sensor data into a surface model of the product is presented. A new surface reconstruction algorithm is proposed to generate the surface model directly from the raw 3D point-cloud data of a product. In addition, an advanced programming algorithm that uses acquired surface models to automatically update robot programmes for a new task with minimal manual intervention is presented.