This project aims to develop novel methods and software tools for the implementation of highly adaptive multi-agent systems for manufacturing control, using plug-and-produce technology. These multi-agent systems will use effectively the available resources (e.g. robots, assembly stations, sensors, measuring machines) to adapt to changing production scenarios and will offer a dramatic reduction in deployment cost and time, and improved operational efficiency and equipment utilisation. This research is strongly related to the Industry 4.0 strategy initiative. The ultimate objective is the creation of "meta-automation" software that will enable reconfiguring automated manufacturing systems and eliminating disturbances with minimum human intervention.
The competitiveness of manufacturing companies in modern markets will increasingly depend on cost-effective flexible automation technologies. The reconfiguration of an automated manufacturing system, for increasing its capacity or introducing a new product variant, requires extensive human intervention and, therefore, it is time-consuming and expensive. Plug-and-produce systems introduced some levels of autonomy. However, there are currently no known plug-and-produce platforms that can address the full cycle of deployment, control and adaptation of heterogeneous mechatronic components, in particular components from different vendors. Despite achievements to date in the fields of multi-agent systems, plug-and-produce technology and flexible manufacturing, fundamental challenges remain. The full potential of the agent-based plug-and-produce approach is still being uncovered, including the ability to handle complex real-time tasks.
The aim of this project is a paradigm shift from a conventional, resource-intensive and largely human-driven configuration and system integration process to plug-and-produce automated manufacturing systems with self-awareness and adaptation capabilities. The key objectives are:
1. To define a vision and software architecture for multi-agent systems for plug-and-produce manufacturing control.
2. To define a plug-and-produce multi-agent approach for the integration of modules from different equipment suppliers and heterogeneous control systems inside one production line.
3. To develop methods for real-time system awareness using integrated sensor networks, and to develop self-learning capabilities for a robust process optimisation of the entire manufacturing system as well as individual production units.
4. To design advanced distributed control infrastructures and scalable cloud architectures enabling resource virtualisation.
5. To design smart human-machine interfaces that can support proactively the user throughout system evolution.