Knee Extension and Flexion Prediction Using Rectus Femoris and Semitendinosus Electromyogram for Controlling Rehabilitation Robot

Title
Knee Extension and Flexion Prediction Using Rectus Femoris and Semitendinosus Electromyogram for Controlling Rehabilitation Robot
Authors
전형진김경재황요하김창환김승종이종민
Issue Date
2013-08
Publisher
The 7th Asian Pacific Conference on Biomechanics
Abstract
Robotic-assisted gait training is able to improve the functional ability of the patients suffering from neurological injuries such as stroke or spinal cord injury. However, in recent studies, robot-based therapy has not shown a distinct enhancement over the conventional physical therapy. To improve the rehabilitation efficacy, applying motor intention of patient to the robot is positively necessary to make him exercise effectively such as an active participation, learning skills and error-drive-learning1. This paper proposes hidden Markov model(HMM)2 of electromyogram(EMG)s as intention recognition method. To verify the suggested method, it was applied to a knee rehabilitation robot and a non-disabled subject. For safety, they were separated from physical contacts. The auto-regressive model parameters of rectus femories(RF, knee extensor) and semitendinosus(ST, knee flexor) EMGs were feature vectors of HMM. The EMG signals of knee extension/flexion/resting states are used to train 3 HMMs. Then the subject straightened and bended his knee, and the suggested method recognized the knee state, and the robot followed the subject’s intention. The results are shown at Figure 1. Upper graphs show the EMGs of 2 muscles. Lower graphs show the joint angles(solid lines) of the knee rehabilitation robot and the recognition results(dashed lines) of the subject’s knee bending/straightening intention. The subject’s knee state was recognized by the suggested method and the robot followed the subject’s intention within 300 msec after EMG onsets (vertical lines). So, the method has the potential to be an intention recognition method.
URI
http://pubs.kist.re.kr/handle/201004/45826
Appears in Collections:
KIST Publication > Conference Paper
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE