Ultrasound Indoor Positioning Dataset

This datasets represents received ultrasonic chirp signals from 14 to 15 speakers in an indoor environment, measured or simulated at 2000 to 7000 different positions in an indoor environment. This dataset contains in total 4 datasets: (I) simulated shoebox, (II) simulated L-shape room, (III) real-world measurements in the Techtile testbed [1], (IV) a synthetic data example of the measurements in the Techtile testbed.

Dataset measured in the Techtile testbed [1] consisting of 15 speakers (anchors). Each speaker transmits an ultrasound chirp signal (20-40 kHz, 30ms). These signals are received at a lot of locations (>2000 for measurements, >7000 for simulation) in order to create a dataset and do indoor positioning. The principles for the model-based indoor positioning system are described in [2]. These datasets are made for research regarding the use of data-driven solutions for ultrasound indoor positioning systems. We tested multiple neural networks (NNs) to find the best performing option. In the end the best NN, outperforming the model-based solution with almost 90% is the graph neural network (GNN). This uses the envelope of the correlated received chirp part and the transmitted chirp in combination with one-hot encoding to encode the anchor to position the mobile node (MN). The paper will later be published by Daan Delabie , Thomas Feys , Chesney Buyle , Bert Cox , Liesbet Van der Perre and Lieven De Strycker [3]. More details can be found in the paper. The information regarding the datasets can be found as readme files inside each directory. The GNN model's and envelope (LPF) data is also given.

[1] D. Delabie, B. Cox, L. De Strycker and L. Van der Perre, "Techtile: a Flexible Testbed for Distributed Acoustic Indoor Positioning and Sensing," 2022 IEEE Sensors Applications Symposium (SAS), Sundsvall, Sweden, 2022, pp. 1-6, doi: 10.1109/SAS54819.2022.9881342.[2] B. Cox, C. Buyle, D. Delabie, L. De Strycker and L. Van der Perre, L., "Positioning Energy-Neutral Devices: Technological Status and Hybrid RF-Acoustic Experiments". (2022). Future Internet, 14(5), 156. https://doi.org/10.3390/fi14050156[3] D. Delabie, T. Feys, C. Buyle, B. Cox, L. Van der Perre and L. De Strycker 'To be published'