[BA] Semantic Modeling and Integration of Real-Time Data Streams Using Feature-Terms

Modern data-intensive systems (e.g. robotics, IoT, or distributed digital twins) must integrate real-time data streams from heterogeneous external sources such as REST APIs, WebSockets, or sensor buses. These sources are often structurally inconsistent or weakly typed, which makes semantic discovery and integration difficult. Feature-Terms \cite{AITKACI1986293} are flexible, algebraic structures used to represent types as labeled feature trees. Unlike rigid schemas or flat records, Feature-Terms support structural subtyping, partial matching, and term rewriting — making them ideal for semantically discovering and integrating diverse data sources by type rather than by fixed identifier or name. The goal of this thesis is to investigate the use of Feature-Terms for modeling real-time stream data. A prototype shall be implemented that allows for the semantic registration, querying, and access of live or virtual streams using a Feature-Term-based interface. The work includes a comparative evaluation against schema-based systems (e.g., GraphQL) and semantic approaches (e.g., RDF/SPARQL).

Betreuer: Sebastian Thielemann