Published November 7, 2025 | Version v1
Dataset Open

FallTL: A Large-Scale Motion Dataset for Pre-Impact Fall Detection based on Wearable Inertial Sensors

Authors/Creators

  • 1. ROR icon Dalian University of Technology
  • 2. ROR icon Northeastern University
  • 3. ROR icon Dalian Polytechnic University
  • 4. ROR icon Universidad Politécnica de Madrid

Description

FallTL is a large-scale motion dataset based on wearable inertial sensors that is suitable for both pre-impact fall detection and post-impact fall detection tasks. It contains data collected from inertial sensors placed on eight body locations, covering 12 types of activities of daily living (ADLs) and 34 types of fall activities. In total, the dataset includes approximately 28.43 hours of motion data. In addition, for each fall activity file, ready-to-use fall labels are provided. Please refer to the README.txt file for more details.

Files

FallTL.zip

Files (403.2 MB)

Name Size
md5:0b0e7e750eac006d466f01aef3589c33
403.2 MB Preview Download