Footwear-Based dataset for Fall Detection

The Footwear-Based dataset for Fall Detection (TST FB4FD) contains data acquired through a pair of smart shoes. The smart shoes are specifically designed for fall detection purposes and are equipped respectively with 3 Force Sensing Resistors (FSRs) and an inertial unit. 

More specifically, the dataset consists of 32 different falls and 8 activities of daily living (ADLs) performed by 17 healthy subjects aged between 21 and 55 years, for a total of 544 falls and 136 ADLs sequences .

 The dataset is also available on IEEE DataPort.

Study Participants and Context

Data acquisition was held in a laboratory environment, involving 17 voluntary users.

As regards the setup, the scene is usually free from any kind of furniture or object. On the ground two foam mats are positioned to soften the blow and protect people from the impact. Only when the specific type of fall requires chairs or armchairs, they have been added to the scene.
 
Below an example of forward fall:

 

Details of participants are reported in the following table:

ID Gender Age Height (cm) Weight (Kg)
ES02 female 40 162 60
ES04 female 29 170 74
ES05 female 25 160 52
ES07 male 30 176 65
ES08 male 55 173 80
ES09 male 21 169 58
ES10 male 21 178 70
ES11 male 23 175 59
ES12 male 28 178 74
ES13 male 28 160 76
ES14 male 26 182 73
ES15 male 40 187 87
ES16 male 21 189 80
ES17 male 22 167 64
ES18 male 22 170 72
ES19 male 21 188 78
ES20 male 21 177 78

Dataset Description

The dataset consists in 17 folders,  named with the user ID to which they refer.

Each folder is structured as follow:

  • Acquisitions:
    • ADLs:
      • APBE: Pick up an object on the floor with bending.
        • APBEST: The subject is standing, bends, picks up an object on the floor, and then stands up again.
        • APBEWK: The subject walks, bends, picks up an object on the floor, and then stands up again.
      • APSQ: Pick up an object on the floor with squatting.
        • APSQST: The subject is standing, squats, picks up an object on the floor, and then stands up again.
        • APSQWK: The subject walks, squats, picks up an object on the floor, and then stands up again.
      • ASCH: Sit and get up from the chair.
        • ASCHST: The subject is standing, sits on a chair, and then stands up again.
        • ASCHWK: The subject walks, sits on a chair, and then stands up again.
      • ASSO: Sit and get up from the couch.
        • ASSOST: The subject is standing, sits on a couch, and then stands up again.
        • ASSOWK: The subject walks, sits on a couch, and then stands up again.
    • Falls: 
      • FBELFR: Backward falls, finishing lying.
        • FBELFRST: The subject is standing, falls backwards, and remains on the ground.
        • FELFRSTRC: The subject is standing, falls backwards, stays on the ground for a while and then gets up again.
        • FBELFRWK:
        • FBELFRWKRC:
      • FBESFR: Backward falls, finishing sitting.
        • FBESFRST: The subject is standing, falls backwards, and remains on the ground.
        • FBESFRSTRC: The subject is standing, falls backwards, stays on the ground for a while and then gets up again.
        • FBESFRWK: The subject walks, falls backward, and remains on the ground.
        • FBESFRWKRC: The subject walks, falls backward, stays on the ground for a while and then gets up again.
      • FFELFR: Forward falls, finishing lying.
        • FFELFRST: The subject is standing, falls forwards, and remains on the ground.
        • FFELFRSTRC: The subject is standing, falls forwards, stays on the ground for a while and then gets up again.
        • FFELFRWK: The subject walks, falls forwards, and remains on the ground.
        • FFELFRWKRC: The subject walks, falls forwards, stays on the ground for a while and then gets up again.
      • FFOKCH: Forward falls on the knees grabbing the chair.
        • FFOKCHST: The subject is standing, falls forwards, and remains on the ground, grabbing the chair.
        • FFOKCHSTRC: The subject is standing, falls forwards, stays on the ground grabbing the chair for a while, and then gets up again.
        • FFOKCHWK: The subject walks, falls backward, and remains on the ground, grabbing the chair.
        • FFOKCHWKRC: The subject walks, falls backward, stays on the ground grabbing the chair for a while, and then gets up again.
      • FFOKFR:Forward falls on the knees.
        • FFOKFRST: The subject is standing, falls forwards, and remains on the ground.
        • FFOKFRSTRC: The subject is standing, falls forwards, stays on the ground for a while and then gets up again.
        • FFOKFRWK: The subject walks, falls forwards, and remains on the ground.
        • FFOKFRWKRC: The subject walks, falls forwards, stays on the ground for a while and then gets up again.
      • FFOKSO: Forward falls on the knees grabbing the sofa.
        • FFOKSOST: The subject is standing, falls forwards, and remains on the ground, grabbing the sofa.
        • FFOKSOSTRC: The subject is standing, falls forwards, stays on the ground grabbing the sofa for a while, and then gets up again.
        • FFOKSOWK: The subject walks, falls forwards, and remains on the ground, grabbing the sofa.
        • FFOKSOWKRC: The subject walks, falls forwards, stays on the ground grabbing the sofa for a while and then gets up again.
      • FSLEFR: Left side falls.
        • FSLEFRST: The subject is standing, falls on the left side, and remains on the ground.
        • FSLEFRSTRC: The subject is standing, falls on the left side, stays on the ground for a while and then gets up again.
        • FSLEFRWK: The subject walks, falls on the left side, and remains on the ground.
        • FSLEFRWKRC: The subject walks, falls on the left side, stays on the ground for a while and then gets up again.
      • FSRIFR: Right side falls.
        • FSRIFRST: The subject is standing, falls on the right side, and remains on the ground.
        • FSRIFRSTRC: The subject is standing, falls on the right side, stays on the ground for a while and then gets up again.
        • FSRIFRWK: The subject walks, falls on the right side, and remains on the ground.
        • FSRIFRWKRC: The subject walks, falls on the right side, stays on the ground for a while and then gets up again.

In each of these sub-folders contains a .txt file, in which each row represents an acquisition. The acquisitions occur every 10 ms for both shoes.

Each line contains the following fields in order:

  • fsr1_l: output voltage value acquired by FSR1, applied on the left heel;
  • fsr2_l: output voltage value acquired by FSR2, applied on the left 1st metatarsal head;
  • fsr3_l: output voltage value acquired by FSR3, applied on the left 5th metatarsal
    head;
  • fsr1_r: output voltage value acquired by FSR1, applied on the right heel;
  • fsr2_r: output voltage value acquired by FSR2, applied on the right 1st metatarsal head;
  • fsr3_r: output voltage value acquired by FSR3, applied on the right 5th metatarsal
    head;
  • accel_x_r: accelerometer reading on the x axis of the right shoe;
  • accel_y_r: accelerometer reading on the y axis of the right shoe;
  • accel_z_r: accelerometer reading on the z axis of the right shoe;
  • mag_x_r: magnetometer reading on the x axis of the right shoe;
  • mag_y_r: magnetometer reading on the y axis of the right shoe;
  • mag_z_r: magnetometer reading on the z axis of the right shoe;
  • gyro_x_r: gyroscope reading on the x axis of the right shoe;
  • gyro_y_r: gyroscope reading on the y axis of the right shoe;
  • gyro_z_r: gyroscope reading on the z axis of the right shoe;
  • accel_x_l: accelerometer reading on the x axis of the left shoe;
  • accel_y_l: accelerometer reading on the y axis of the left shoe;
  • accel_z_l: accelerometer reading on the z axis of the left shoe;
  • mag_x_l: magnetometer reading on the x axis of the left shoe;
  • mag_y_l: magnetometer reading on the y axis of the left shoe;
  • mag_z_l: magnetometer reading on the z axis of the left shoe;
  • gyro_x_l: gyroscope reading on the x axis of the left shoe;
  • gyro_y_l: gyroscope reading on the y axis of the left shoe;
  • gyro_z_l: gyroscope reading on the z axis of the left shoe.

Citation

Cite this dataset as:

IEEE

Susanna Spinsante, Ennio Gambi, Laura Montanini, Davide Perla, Antonio Del Campo, "TST Footwear-Based dataset for Fall Detection (TST FB4FD)", IEEE Dataport, 2017. [Online]. Available: http://dx.doi.org/10.21227/H2W01S. Accessed: Apr. 11, 2017. 

APA

Susanna Spinsante, Ennio Gambi, Laura Montanini, Davide Perla, Antonio Del Campo. (2017). TST Footwear-Based dataset for Fall Detection (TST FB4FD). IEEE Dataport. http://dx.doi.org/10.21227/H2W01S

MLA

Susanna Spinsante, Ennio Gambi, Laura Montanini, Davide Perla, Antonio Del Campo. (2017). "TST Footwear-Based dataset for Fall Detection (TST FB4FD)." Web.

BIBTEX

@data{h2w01s-17,
doi = {10.21227/H2W01S},
url = {http://dx.doi.org/10.21227/H2W01S},
author = {Susanna Spinsante; Ennio Gambi; Laura Montanini; Davide Perla; Antonio Del Campo },
publisher = {IEEE Dataport},
title = {TST Footwear-Based dataset for Fall Detection (TST FB4FD)},
year = {2017} }

Download

Click this link to download the TST FB4FD dataset: http://www.tlc.dii.univpm.it/fb4fd/

Contacts

For more information, please contact the authors: 
s.spinsante@univpm.it
e.gambi@univpm.it
laura.montanini@univpm.it
d.perla@pm.univpm.it
a.delcampo@pm.univpm.it

Comments are closed