Motion Learning for Emotional Interaction and Imitation of Children with Autism Spectrum Disorder
2017
Conference Paper
hi
We aim to use motion learning to teach a robot to imitate people's unique gestures. Our robot, ROBOTIS-OP2, can ultimately use imitation to practice social skills with children with autism. In this abstract, two methods of motion learning were compared: Dynamic motion primitives with least squares (DMP with WLS), and Dynamic motion primitives with a Gaussian Mixture Regression (DMP with GMR). Movements with sharp turns were most accurately reproduced using DMP with GMR. Additionally, more states are required to accurately recreate more complex gestures.
Award: | (First place tie in category, "Biomedical Engineering, Graduate Research") |
Author(s): | Rachael Burns and Sam Cowin |
Book Title: | 2017 GW Research Days, Department of Biomedical Engineering Posters and Presentations |
Year: | 2017 |
Month: | April |
Department(s): | Haptic Intelligence |
Bibtex Type: | Conference Paper (inproceedings) |
Paper Type: | Abstract |
Event Place: | Washington, DC, USA |
Award Paper: | First place tie in category, "Biomedical Engineering, Graduate Research" |
Language: | English |
URL: | https://hsrc.himmelfarb.gwu.edu/gw_research_days/2017/BME/17/ |
Attachments: |
|
BibTex @inproceedings{Burns17-GWRD-MotionLearning, title = {Motion Learning for Emotional Interaction and Imitation of Children with Autism Spectrum Disorder}, author = {Burns, Rachael and Cowin, Sam}, booktitle = {2017 GW Research Days, Department of Biomedical Engineering Posters and Presentations}, month = apr, year = {2017}, doi = {}, url = {https://hsrc.himmelfarb.gwu.edu/gw_research_days/2017/BME/17/}, month_numeric = {4} } |