Program and operate robots without expert knowledge
The state of the art in production using industrial robots is mainly concentrated on large scale production facilities. Small and medium-sized enterprises (SMEs) are mostly reluctant to consider automation using robots. Unlike large scale industries, SMEs have small slot size and constantly changing product cycles. These facts make the use of typical industrial robot work cells difficult since, they are not easy to program and require skilled technicians for operation. Furthermore, SME shop floors are less structured as compared to their counterparts in large industrial settings, making robot deployment even more challenging.
The goal of the SMErobotics project is to bring industrial robots to SME production lines. For this purpose, the project aims towards developing a robot system which supports intuitive programming and can easily adapt to changing product cycles. Most importantly, this robot system should be able to cooperate with an average skilled human worker. Special attention will be given to the system error recovery by means of assistance from the human worker, robot system or both.
From this project, the SMEs will immensely benefit in the form of continuous improvement in throughput, human working standards and production quality. Moreover, with such a flexible robot system, they could easily adapt to their changing product cycles while still maintaining quality standards, without the need of expert knowledge. This results in reduced installation and operation costs.
During the course of this project, systems in the area of artificial Intelligence will be researched and developed. This will be a contribution to the state of the art in symbiotic interaction and mutual learning.
In the SMErobotics project, fortiss will be responsible to deliver system modules for multi-modal sensor fusion/fission, perception, reasoning and planning. This task will be achieved by using, extending and implementing state of the art tools/techniques of artificial intelligence (reasoning/planning) on standard industrial robots. These modules will be integrated together with modules for process models, task execution under uncertainties, and robot control, which will be developed by our project partners.