IN-Rep: A New Open Data Repository for AI-Based Positioning in Industrial Networks

As the adoption of wireless networks in indus-trial premises rises, the availability of experimental data is of paramount importance to enable the design of reliable commu-nications and positioning systems, capable of operating under the harsh propagation conditions that characterize industrial environments by leveraging recent advances in the application of Artificial Intelligence and Machine Learning techniques. This work focuses in particular on positioning, and reviews recent con-tributions relying on experimental data, assessing the availability of such data. The review highlights that Ultra Wide Band and Visible Light Communications are by far the two most popular technologies for positioning in industrial settings, but also that experimental data are seldom made available to the research community, and typically for a single technology, This work fills this gap by introducing IN-Rep: an open data repository for datasets collected in industrial environments. IN-Rep is planned as an ongoing initiative, to be populated over time with multiple datasets in different propagation environments and combining multiple technologies. Three UWB datasets are contextually released: two preexisting but not previously available, and one newly collected. The new dataset, including both Channel Impulse Responses and distance measurements between transceiver pairs collected in an industrial center, is then described, and its relevance in supporting recent efforts on the application of Artificial Intelligence and Machine Learning techniques to positioning is finally discussed.