TY - GEN
T1 - Non-contact capacitance sensing for continuous locomotion mode recognition
T2 - 2013 IEEE 13th International Conference on Rehabilitation Robotics, ICORR 2013
AU - Zheng, Enhao
AU - Wang, Long
AU - Luo, Yimin
AU - Wei, Kunlin
AU - Wang, Qining
PY - 2013
Y1 - 2013
N2 - Locomotion mode recognition plays an important role in the control of powered lower-limb prostheses. In this paper, we present a non-contact capacitance sensing system (C-Sens) to measure the interfacial signals between the residual limb and the prosthetic socket. The system includes sensing front-ends, a sensing circuit, a control circuit and foot pressure insoles. In the proposed system, the electrodes are fixed on the inner surface of the socket, which couple with the human body forming capacitors. The foot pressure insoles are built for detecting gait phases. The data sequence is controlled by the control circuit. To evaluate the capacitance sensing system, experiments with a transtibial amputee are carried out and seven kinds of locomotion modes are recorded. With the continuous phase dependent classification method and the quadratic discriminant analysis (QDA) classifier, the average recognition accuracies are 93.8% and 95.0% for the stance phase and the swing phase respectively. The results show the potential of the proposed system for the control of powered lower-limb prostheses.
AB - Locomotion mode recognition plays an important role in the control of powered lower-limb prostheses. In this paper, we present a non-contact capacitance sensing system (C-Sens) to measure the interfacial signals between the residual limb and the prosthetic socket. The system includes sensing front-ends, a sensing circuit, a control circuit and foot pressure insoles. In the proposed system, the electrodes are fixed on the inner surface of the socket, which couple with the human body forming capacitors. The foot pressure insoles are built for detecting gait phases. The data sequence is controlled by the control circuit. To evaluate the capacitance sensing system, experiments with a transtibial amputee are carried out and seven kinds of locomotion modes are recorded. With the continuous phase dependent classification method and the quadratic discriminant analysis (QDA) classifier, the average recognition accuracies are 93.8% and 95.0% for the stance phase and the swing phase respectively. The results show the potential of the proposed system for the control of powered lower-limb prostheses.
UR - https://www.scopus.com/pages/publications/84891112871
U2 - 10.1109/ICORR.2013.6650410
DO - 10.1109/ICORR.2013.6650410
M3 - 会议稿件
C2 - 24187229
AN - SCOPUS:84891112871
SN - 9781467360241
T3 - IEEE International Conference on Rehabilitation Robotics
BT - 2013 IEEE 13th International Conference on Rehabilitation Robotics, ICORR 2013
Y2 - 24 June 2013 through 26 June 2013
ER -