Multi-Person Localization and Vital Sign Estimation Radar Dataset

Citation Author(s):
Christian A.
Technische Universität Darmstadt
Technische Universität Darmstadt
Technische Universität Darmstadt
von Stryk
Technische Universität Darmstadt
Abdelhak M.
Technische Universität Darmstadt
Technische Universität Darmstadt
Submitted by:
Christian Schroth
Last updated:
Thu, 06/15/2023 - 03:32
Data Format:
1 rating - Please login to submit your rating.


The large number and scale of natural and man-made disasters have led to an urgent demand for technologies that enhance the safety and efficiency of search and rescue teams. Semi-autonomous rescue robots are beneficial, especially when searching inaccessible terrains, or dangerous environments, such as collapsed infrastructures. For search and rescue missions in degraded visual conditions or non-line of sight scenarios, radar-based approaches may contribute to acquire valuable, and otherwise unavailable information.

This dataset provides the possibility to develop algorithms for, e.g., radar-based (through-wall) multi-person detection, localization, 3D direction-of-arrival estimation, breathing frequency estimation or heart beat estimation. The challenging dataset was collected using a semi-autonomous robot equipped with a commercially available through-wall radar system. The dataset is composed of 62 scenarios of various difficulty levels with up to five persons captured in different postures, angles and ranges including wooden and stone obstacles that block the radar line of sight. Ground truth data for reference locations, respiration, electrocardiogram, and acceleration signals are included.

The ethics commission of 'Technische Universität Darmstadt' (EK 30/2023, date of approval: 10. May 2023) gave their consent in recording and publication of the dataset.


The dataset is composed of 121 files with 62 radar measurements and 59 reference vital parameter measurements. The radar measurements with IDs 0, 0-1 and 0-2 have no corresponding reference file as they are empty background measurements. Each measurement containing a person has a length of 2000 slow time samples (i.e., about 160 s), while the empty room settings have a length of 200 slow time samples (i.e., about 16 s). An overview of the measurements is given in Table III and a detailed list of all measurements containing the type of obstacle, and the positions and postures of the subjects is provided in the supplementary material.

Positions of the participants were chosen in a way to ensure different levels of difficulties. Recording IDs M1 to M11 contain measurements of a single participant with different postures, angles, ranges and obstacles. The next set of measurements with IDs M12 to M26, are the most challenging free space measurements as they contain measurements of five or four participants in different ranges and angles, including multiple participants in the same range. Starting with measurement ID M27 all measurements are performed three times, for free space, with wall and with door. This set contains measurements with two and three participants in various arrangements. Scenarios in which one person is blocking the line of sight of another person were not recorded.

More details can be found in the accompanying pdf documentation.


Funding Agency: 
LOEWE-Zentrum emergenCITY

Dataset Files

Open Access dataset files are accessible to all logged in  users. Don't have a login?  Create a free IEEE account.  IEEE Membership is not required.