[BA] Development and Evaluation of a Semantic Middleware for Dynamic Stream Discovery in Heterogeneous Environments based on Feature Terms

Modern industrial IoT and robotics environments generate highly dynamic data streams across various protocols, and while prior work has demonstrated that Feature Terms offer a powerful abstraction for semantically describing and discovering these streams via a domain-specific language (DSL), these approaches remain isolated software components without integration into industrial standard protocols and have not been evaluated under realistic conditions. This thesis aims to combine, extend, and evaluate these prior approaches into an integrated middleware system capable of capturing real data sources, such as sensors via MQTT or robot nodes via ROS 2, at runtime, semantically describing them, registering them, and exposing them to consumers via the DSL. A core contribution is the independent development and application of an evaluation framework to assess the practical feasibility of Feature-Term-based stream discovery both quantitatively and qualitatively under conditions such as high volatility or large sensor counts.

Betreuer: Sebastian Thielemann