TY - JOUR
T1 - Plastic Optical Fiber Enabled Smart Glove for Machine Learning-Based Gesture Recognition
AU - Li, Jie
AU - Liu, Bin
AU - Hu, Yingying
AU - Liu, Juan
AU - He, Xing-Dao
AU - Yuan, Jinhui
AU - Wu, Qiang
N1 - Funding information: This work was jointly supported by National Natural Science Foundation of China (NSFC) (11864025, 62175097 and 62065013); Natural Science Foundation of Jiangxi Province (Grant No. 20212BAB202024); Royal Society International Exchanges 2021 Round 2 (IES\R2\212135).
PY - 2024/4/1
Y1 - 2024/4/1
N2 - Gesture recognition has always been an important research direction in the field of human-computer interaction (HCI). In this paper, a wearable gesture recognition system basedon D-shaped plastic optical fiber (POF) curvature sensor was proposed and experimentally studied. A highly bend sensitive Dshaped POF curvature sensor was made and integrated into a fivechannel signal acquisition system on a PCB board (8×4.5 cm), which was embedded into an elastic glove to collect fingers’ movement data. Thirteen gestures and eleven grasping actions were defined, and the gesture data, the grasping action data and the gesture data mixed with grasping action data were normalized, calibrated and imported into a support vector machine (SVM) classifier based on Gaussian kernel function and feedforward neural networks (FNN) respectively. The recognition accuracy based on SVM of 13 gestures and 11 grasping actions reached 99.8% and 97.7% respectively. The recognition accuracy of 13 kinds of gesture data mixed with 11 kinds of grasping action data based on Gaussian kernel function in SVM classification model and FNN were 98.9% and 99.4% respectively.
AB - Gesture recognition has always been an important research direction in the field of human-computer interaction (HCI). In this paper, a wearable gesture recognition system basedon D-shaped plastic optical fiber (POF) curvature sensor was proposed and experimentally studied. A highly bend sensitive Dshaped POF curvature sensor was made and integrated into a fivechannel signal acquisition system on a PCB board (8×4.5 cm), which was embedded into an elastic glove to collect fingers’ movement data. Thirteen gestures and eleven grasping actions were defined, and the gesture data, the grasping action data and the gesture data mixed with grasping action data were normalized, calibrated and imported into a support vector machine (SVM) classifier based on Gaussian kernel function and feedforward neural networks (FNN) respectively. The recognition accuracy based on SVM of 13 gestures and 11 grasping actions reached 99.8% and 97.7% respectively. The recognition accuracy of 13 kinds of gesture data mixed with 11 kinds of grasping action data based on Gaussian kernel function in SVM classification model and FNN were 98.9% and 99.4% respectively.
KW - Bending
KW - D-shaped POF curvature sensor
KW - FNN
KW - Gaussian kernel function
KW - Gesture recognition
KW - Optical fiber amplifiers
KW - Optical fiber sensors
KW - Optical fibers
KW - SVM
KW - Sensors
KW - Support vector machines
KW - human-computer interaction (HCI)
UR - http://www.scopus.com/inward/record.url?scp=85161037775&partnerID=8YFLogxK
U2 - 10.1109/tie.2023.3277119
DO - 10.1109/tie.2023.3277119
M3 - Article
SN - 0278-0046
VL - 71
SP - 4252
EP - 4261
JO - IEEE Transactions on Industrial Electronics
JF - IEEE Transactions on Industrial Electronics
IS - 4
ER -