The manufacturing industry is facing strong challenges related to an ageing and increasingly scarce workforce and an increasing demand for small or medium series highly-customised complex products. Robot manufacturers and other high-tech companies responding to this demand, have invested and still are investing heavily in new technologies and have introduced two-armed robots, mobile manipulators, and other “multi-robot” systems to the market.
Although there is a lot of “buzz” on “flexible”, “collaborative” and “advanced” robotics, to address the aforementioned challenges, real industrial uptake of these systems is still limited today to a small number of early adopters and to factory proof-of-concepts. A key technological barrier impeding widespread use in the Flemish manufacturing industry (and beyond) is the difficulty to deploy and use complex multi-robot systems in an efficient way in real-life industrial environments using state-of-the-art robot programming approaches. The software limitations of these systems have a vehemently negative impact on the return on investment.
The objective of this project is to build an integrated software framework to remove this technological barrier. The framework focuses on providing an integrated solution for three coupled topics:
- optimal continuous-time motion planning and control
- a domain-specific language and negotiation mechanism for task coordination and discrete decision making
- a generic world model description, including a query language and sensor fusion techniques. For each of the topics both algorithms and software support will be developed and integrated into an overall framework
Evidently, the framework can also be applied to (non-collaborative) industrial robots, and for some companies this will be the first valorisation phase in follow-up ICONS/R&D/B2B activities. In order to avoid the possibility that the framework quickly becomes deprecated when scaling up to more complex systems of robotic subsystems, it is however crucial that multiple robot subsystems are considered right from the start in the framework design phase. Currently, multiple robot manufacturers are facing strong software legacy issues when developing redundant (e.g. seven-axis robot arms) or two-armed robot systems, because their software was designed for non-redundant six-axis robot manipulators. This legacy resulted in a lot of ad hoc extension modules, that are very case-specific, not user friendly and not transferrable from one robot platform to another.
The project will develop demonstrators for machine tending, assembly, and contour tracing applications, and is backed by a strong user group that sees the strategic value of the framework for the coming five to ten years, consisting of
- high-tech companies, service providers, system integrators and robot manufacturers that will pick up the framework: Achilles Design, FMI Industrial Modules, FRS, Intermodalics, KUKA, Octinion, Optidrive, and Pick-it;
- manufacturing companies that will integrate these systems in their factories (directly or through an integrator): Atlas Copco, Audi Brussels, Bekaert, CNH Industrial, Niko, Terumo, Van Hool, and Vitalo.