CORDIS Archive

View the original page arrowbar Legal Noticebar Print the page
This page has been archived. It will no longer be updated.
[next stand][Exhibition home page][EITC 97 home page]

RETRARO

Intelligent robot avoids obstacles

Real-time expert systems technology helps robots plan and navigate their way around building obstacles.

Robots are ideal for transporting goods around factories. To date however, most installed AGVs (Autonomous Guided Vehicles) have relied on induction wires for guidance. Such robots therefore depend totally on the physical configuration of wires within the building. The resulting lack of flexibility is a real problem for plant managers wanting to move or purchase production equipment. Moreover, time and money constraints often oblige companies to rely on traditional methods, due to the complexity of preparing an AGV's configuration.

Under the RETRARO project, a consortium is developing an operations system for AGVs transporting materials in a textile production unit. This software application uses real-time expert system technology mounted on an on-board computer. The project also calls on the REAKT platform (the fruit of two Esprit projects), which includes a software environment, methodology and tool kit for developing general purpose real-time knowledge-based control systems.

The robot's free movement - it is not guided by induction wires - makes it highly flexible and able to move around a factory almost without hindrance. An array of low-cost sensors, which provide data on the AGV's position and the presence of obstacles, allow it to choose the best way to perform a task. Any changes in building layout or production lines may be accommodated simply by updating the robot's configuration file. Trials at Schoeller Bregenz in Germany have already shown that the AGV improves the production process; increased flexibility and reliability lead to cost reductions and greater competitiveness. The intelligent control system implemented in this robot can also be applied to other AGV models. Further work will concentrate on the dynamic planning of tasks, and AGV positioning and obstacle detection.

Several features of the project's AGV - the user interface with the path selection and replanning capabilities - are demonstrated here on a PC, which is connected to an intelligent mini-robot on a small platform.

Contact: Pedro LÓPEZ PEÑA, GMV SA (E)
tel +34 1 807 2196 - fax +34 1 807 2199 - plopez@gmv.es - www.gmv.es/


EITC 97 home page | Exhibition home page

This document is located at /esprit/src/exhib43.htm
It was last updated on 22 October 1997, and is maintained by the EITC 97 Secretariat - eitc97@dg3.cec.be