direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

MA: Anomaly Detection Agent for Industry 4.0

Digitalization drives the current Industry processes. Data is collected on several different sources, creating large sets of data, also called as Big Data. Such sensors, which produce the data, collect the data at the edge and send through a gateway to the cloud. As not all collected data is relevant for the production process, automatic solutions need to be applied to detect and present the most relevant data. Furthermore, the data appears as a data stream and should be also analyzed as such. Therefore, we created anomaly detection algorithms, which can be applied on data streams for black box services.  

In this thesis, the student should create an anomaly detection agent, which can be deployed on an edge device. This service should communicate with a management component, which can visualize results, but also is able to configure agents in the edge. Also, the user should be able to retrain on historic data models, which can be sent to the edge, to provide even better detection results by expert based models.

In this thesis, the following tasks are therefore included:

  • Developing of an overall architecture for the anomaly detection Industry 4.0 use case, using hyped frameworks like e.g. Kafka, Flink, Spark, ElasticSearch, etc. (provide extensive survey) 
  • Creation of an anomaly detection agent, handling the algorithms developed at our department
  • Development of a management component handling several anomaly detection agents
  • Component to retrain anomaly detection models on historic data and possibility to update agents
  • Visualization component for anomaly detection results of given agents (usage of e.g. Grafana, etc)

Prerequisites for working on this topic are advanced knowledge in algorithmic design and good programming skills in either Java, Go or Python.

Zusatzinformationen / Extras

Quick Access:

Schnellnavigation zur Seite über Nummerneingabe

Auxiliary Functions


Florian Schmidt
+49 (30) 314 28306
Room E-N 101