- To develop cognitive layers that will support scalability and evolution of large-scale heterogeneous IoT-based systems and autonomously manage changes after deployment. COSIBAS will extend the benefits of the IoT concept to industrial applications, in order to enable interoperability for legacy systems by integrating legacy components and industrial systems into existing IoT platforms and by developing automatic driver generation tools that will facilitate integration of other legacy (and future) systems.
- To deliver intelligence and awareness for enhanced business applications and systems through cognitive enablers and services. On the higher level of the software stack, there will be cognitive services like decision management, efficient data analysis and inference, and cognitive assistants with adaptive user-specific interfaces.
- To create specific reasoning modules that will make use, among other things, of the inferential capabilities associated with USM as part of the cognitive services. These modules will enable self-organisation in order to provide continuous auto-commissioning capabilities to IoT systems and services, fault forecasting to predict maintenance schedules of IoT devices and platforms, dependability for timely delivering the required quality of service during the entire system’s life-cycle and self-adaptation to improve system performance during context changes.
- Enable industrial legacy systems with state-of-the-art cognitive tools and communication mechanisms to modernise infrastructures and investments extending the effective life of IT systems while reutilising assets. The application of COSIBAS results and its integration in real scenarios will allow to demonstrate that with the application of new techniques and methods is possible to save OPEX without the need of extraordinary investments or completely migrating or replacing well-functioning systems. This also will reduce the rejection of staff (aversion to technological changes due to the threat of substitution) and will decrease significantly the learning curve.
