DatasetinthePaper    2022 No.2

Chief-editor of dataset: Mi He

The K band ultrawideband radar spectrogram dataset for fall detection 1.0(K-UWB-MSHA-1.0)releases 3240 cases of human behavior data of 36 subjects using UWB radar in two scenes. The K-band UWB radar system operated with the frequency range of 22.8~24.8 GHz and transmited and received the frequency modulated continuous wave signal waveform.

This dataset includes radar data of activities of daily life (ADL) and falls from 36 subjects. Action types include 20 kinds of ADL, such as walking, jogging, sitting down, lying down, standing up, going up and down stairs slowly, walking on crutches, jogging and crawling, and 10 kinds of fall activities on a sponge mat, such as fainting, falling, sliping, tripping up and down stairs in different directions. The size and scene layout of the two rooms for data collection are different (details are seen in the paper). The corresponding protocol was approved by the ethics committee of Army Medical University with approval number AF/sc-03/1.0, 2021. Data of 30 subjects in Scene 1 were used as the training set, and the data of 6 subjects in Scene2 were used as the testing set. The duration of the radar spectrogram data segment is 4 seconds, and the corresponding data preprocessing method is shown in the paper. 

The instruction for this dataset presents the detailed information of data annotation and action types.

Citation formatHE Mi, PING Qinwen, DAI Ran. Research on Fall Detection Based on Deep Learning Fusing Ultrawideband Radar Spectrograms[J]. Journal of Radars, in press. doi: 10.12000/JR21193

Download 874.21MB(4)

Data Usage Protocol of Journal of Radars

The data can be used free of charge for scientific research, teaching and so on, but the data source should be marked in the reference according to the citation format.

Data use for commercial purposes requires permission from the Editorial Department of Journal of Radars.