Ortelio is leading an EU-wide research initiative to help increase the maturity of robotic technologies and enable translation to real applications in real open world environments.
We believe that real applications targeting open world environments need to deal with and reason about the uncertainties and unstructured nature of both the application and the environment. Reasoning on large, uncertain environments requires extensive knowledge bases and resources. Additionally, the gap between symbolic and sub-symbolic reasoning needs to be addressed and closed.
While, for instance, deep learning does a good job at object recognition, it lacks the semantics/context of the task and domain at hand. We expect that by combining the outcomes of sub-symbolic algorithms such as deep learning with semantic reasoning (what we call the ubiquitous semantic approach), the strength of both worlds can be leveraged and higher level of technological robustness achieved.