This dataset consists of IQ samples gathered at Ghent University's Industrial Internet of Things (IIoT) laboratory. It contains timestamped experimental data from 39 experimental scenarios, split into five distinct categories: the tag is in continuous motion, the tag is in continuous motion with obstacles between the tag and anchor, the tag is in continuous motion with no obstacles but the distance between the tag and anchor is constantly varying, the tag halts every 5 degrees for 25 seconds, and the tag halts every 5 degrees for 25 seconds with obstacles between the tag and anchor. The position of the tag has been constantly monitored by a Motion Capture system and measured to a millimetre-level accuracy. This can be used as-is as the Ground Truth (GT) label for single-anchor localization or converted into an azimuth (or even elevation) angle for use in improving on Direction of Arrival determination. More information on the data collection process can be found in the associated paper. File Structure: The recommended Folder structure for the downloaded dataset consists of: Dataset_Root\ | +---->DatasetFiles\ | | | +----->Conversions\ | | | | | +----->TI.json | | | +----->Data\ | | | | | +-------->... | | | +----->metadata.pdf | | +---->Code\ | | | +--------->config.json | | | +--------->requirements.txt | | | +--------->Scripts\ | | | +------->... | +---->Cache\ | | | +--------->TF\ | | | +---->Figures\ | +---->Models\ Metadata: the file metadata.pdf contains a table listing the experiment variables, the number of data points in that experiment, the length of the experiment in seconds, as well as the date and time that the experiment was performed. Instructions: Instructions on how to use the dataset and the scripts that come with it can be found in Instructions.txt

One significant challenge in research is to collect a large amount of data and learn the underlying relationship between the input and the output variables. This data, collected in a laboratory setting, is intended to approximate real-world industrial scenarios. The samples were collected using a Texas Instruments (TI) BOOST-XL-AOA antenna array and the Ground Truth (GT) position of the tag was tracked and logged by a motion capture system with millimetre accuracy. This position was used to calculate the angle between the tag and the antenna array. The collected samples and the process of generating GT labels were validated using the TI phase difference of arrival (PDoA) implementation on the data, yielding a mean absolute error (MAE) at one of the heights without obstacles of 25.71 degrees.