Integration Of Machine Vision In Industrial Robot For Automated Composite Panel Punching
Loading...
Date
2018-11-01
Authors
Mohamed Hanafee, Ameer Mohamed Abdeel Aziz
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
This thesis employs visual servoing control to provide accurate information for
repetitive task on the known object. The problem statement on this thesis involves two
main points. Firstly, by using the conventional method it is not possible to have one
workstation to machine different pattern of composite panels, nor using different
fixtures and tools at the same time. Secondly, the current KUKA Robot Controller
(KRC) 4 does not have the capability to communicate with an external device due to
the controller safety protocol. The aim of this research is to build a machine vision
algorithm for assisting KUKA robot in recognizing and calibrating various shapes of
composite panels precisely. The second objective is to develop and integrate
communication interface program for KRC4 and Raspberry Pi (RPi) microcontroller
and to evaluate the performance of visual servoing control system for automated
composite panel punching. By using socket module for Python programming
language, the feedback control system is built by communicating RPi with KRC4
through Ethernet using KUKA Robot Language (KRL) to enable the robot to execute
the whole process. The image are traced using the machine vision function. Shape
factor value of each four different panels sorted through the graph for 15 process cycle.
Position and rotation of panels are compared to three coordinate systems which are
camera, workpiece and machining center. The differences of these coordinates with
the reference coordinate translated as error and adjustment was made to remove the
error. As a conclusion, KUKA robot integration with RPi enables the machining
process of different panels executed in a workstation only.