The reliable detection of transitions between appliance states in multivariate sensor signals is fundamental to the adoption of Machine Learning methods in industrial IoT. As state transition detection allows a discretization of appliance states, it precedes downstream categorization and integration. Thus, the detection of these transition events is key towards improved manufacturing decision intelligence solutions.
During this internship, the student will assess and extract a suitable set of event detection methods to cover typical sensor readings in iIoT. Next, she will implement these methods within a real-time streaming architecture. Along with that, requirements for a user interface approach to select appropriate methods in a semi supervised way shall be addressed.
Technologies: SQL, timescaledb, R, python, azure.