Handwriting and posture quality measures of children while interacting with a social robot in handwriting activities

About Dataset

This dataset is regarding posture quality measures and handwriting quality measures of children (around 10 y.o.) while interacting with a social robot for about 30 minutes.

About Data collection methodology

The interaction between the robot and the participant was designed combining the phases of by following sequence:1. The greeting phase2. A previous handwriting evaluation (HWE1)3. The first training phase4. An intermediate handwriting evaluation (HWE2)5. The second training phase6. A last handwriting evaluation (HWE3)7. The goodbye phase at the end.

The minimum time for the training phases (steps 3 and 5) was set to 5 minutes. This means that after 5 minutes of training, instead of suggesting the next game, the robot guided the student to the next phase, which in this case was always a HWE.

The data of this dataset is the posture and handwriting measures collected during the steps 2, 4, and 6 of the interaction.

The handwriting measures were collected using the Dynamilis app, whereas the posture were coded by the authors to provide the measures of Rapid entire body assessment (REBA), and Rapid Upper Limb Assessment (RULA). 

The handwriting measure is based on the text they wrote during the evaluation phases while the posture measures are the average score of all the corresponding readings of that specific measure during the period children took to perform the text writing.

Description of the data

Data structure:

Root Dir/  -HandwritingEvaluation/    -ID01.csv    -ID02.csv    -...    -ID031.csv  -ergo_analysis_all_subs_complementary.csv  -README.md

The file "ergo_analysis_all_subs_complementary.csv" contains the posture measures, where the column "sub_id" is the Id of the subject, the column "type" is the phase evaluated (being "A" and "B" for pre- and post- test, respectively, and “1” for after the instructor correcting the subject's posture). All the other columns are the corresponding average of REBA or RULA measures (according to the column's name) read by our system. Names of the columns follow the pattern of "_", for instance the column "REBA_neck_score" tells that the number there is the neck score average using the REBA method.

The handwriting evaluation measures are stored in the files inside the directory "HandwritingEvaluation", where each file contains the 3 handwriting measures of the subject indicated in the file's name. The rows' order is related to their order in the experiment, so row 1 corresponds to the first measure ("A" type in posture measure files), the second row corresponds to the second evaluation phase ("B" in posture) and so on.

Authors

Tozadore, Daniel. 

Wang, Chenyang.

Bruno, Barbara.

Dillenbourg, Pierre.

All authors were with the EPFL at the time of this data collection.

License

This project is licensed under the MIT License - see the [LICENSE.md](LICENSE.md) file for details

Acknowledgments

We acknowledge the SNF for funding this project under the registration FNS 200021E_189475/1.