Assessing human-human therapy kinematics for retargeting to human-robot therapy
2015
Conference Paper
hi
In this paper, we present experiments examining the accuracy of data collected from a Kinect sensor for capturing close interactive actions of a therapist with a patient during stroke rehabilitation. Our long-term goal is to map human-human interactions such as these patient-therapist ones onto human-robot interactions. In many robot interaction scenarios, the robot does not mimic interaction between two or more humans, which is a major part of stroke therapy. The Kinect works for functional tasks such as a reaching task where the interaction to be retargeted by the robot is minimal to none; though this data is not good for a functional task involving touching another person. We demonstrate that the noisy data from Kinect does not produce a system robust enough to be for remapping to a humanoid robot a therapit's movements when in contact with a person.
Author(s): | Michelle J Johnson and Seethu M Christopher and Mayumi Mohan and Rochelle Mendonca |
Book Title: | Proceedings of the IEEE International Conference on Rehabilitation Robotics (ICORR) |
Year: | 2015 |
Month: | August |
Department(s): | Haptische Intelligenz |
Bibtex Type: | Conference Paper (inproceedings) |
Paper Type: | Conference |
DOI: | 10.1109/ICORR.2015.7281312 |
Address: | Singapore |
BibTex @inproceedings{Johnson15-ICORR-Therapy, title = {Assessing human-human therapy kinematics for retargeting to human-robot therapy}, author = {Johnson, Michelle J and Christopher, Seethu M and Mohan, Mayumi and Mendonca, Rochelle}, booktitle = {Proceedings of the IEEE International Conference on Rehabilitation Robotics (ICORR)}, address = {Singapore}, month = aug, year = {2015}, doi = {10.1109/ICORR.2015.7281312}, month_numeric = {8} } |