IoT sensors are everywhere. It is a lot of manual work to label sensors with type and location, but this met information is needed in order for the data to be useful.
The problem we would like to solve: Given a set of sensors at different locations, can we identify which of the sensors are of the same type and can we identify which sensors are located where based on the time-series data alone.
We expect to develop a deep neural network that can perform the relation inference.
The project is done in collaboration with TrønderEnergi, and we will use SCADA data from hydro plants.