RIC: RAPP Improvement Center – UML Diagram
During the last weeks, CERTH/ITI has mainly focused in RAPP’s technical aspects and specifically in RIC’s (RAPP Improvement Center) internal structure and capabilities. A draft UML file of our initial implementation follows:
RIC consists of a number of modules:
- A central MySQL database, where all the personalized information will be stored. By “personalized information” we mean the users’ profiles, which robots they possess and which RApps (Robotic Applications) they have used.
- A centralized knowledge repository where different concepts are stored in an ontology scheme. Since the RAPP project is multidisciplinary, we intend to use the KnowRob robotic ontology, in conjunction with the OpenAAL assisted living ontology. There, anonymous data will be stored, whose relation to the users will be maintained in the Database.
Heterogeneous HOP services that will provide access to heavy duty algorithms, such as Image processing, Speech to text recognition and others. Additionally, such services will be responsible for fetching, storing and returning data from / to the users. The HOP services can be accessed via authentication and can be utilized by all Robotic Applications.
- Finally, RIC will include machine learning algorithms that will produce new knowledge from the existing one. Some examples are the creation of general object models for their visual recognition, detecting patterns in RApps and their employment by the users and others.
CERTH/ITI’s short term plans is to move forward with a premature RIC implementation, with the collaboration of ORTELIO regarding the RAPP Store and some AI functionalities, as well as WUT for the construction of robotic algorithms.