BROWSE
Detail
Conference Paper
Lower body action classification using unlabeled predicted motion
Cited 0 time in 
Share 
Share
- Authors
- Ho Chul Shin, Dong-Woo Lee, Son Yong Ki, Hur Ki Soo
- Issue Date
- 2023-10
- Citation
- International Conference on Information and Communication Technology Convergence (ICTC) 2023, pp.1-3
- Publisher
- IEEE
- Language
- English
- Type
- Conference Paper
- Abstract
- In this study, a rule-based motion classification method was presented for the lower extremity motions prediction and classification for exoskeleton robot lower extremity support. With the CMU public DB and additional motion DB using Xsens device, the joint angles of the lower extremity for the next second was predicted using the bidirectional LSTM algorithm. By defining the motion vector of the predicted lower extremity joint angle and calculating the motion score, future lower extremity motions could be classified as walking, squat, and stoop without labeling process. The proposed algorithm is fast in calculation and does not require label work, so it can be easily loaded into low-cost devices, so it is expected that it can be applied not only to exoskeleton robot control but also to wearable smart devices.
- KSP Keywords
- Action Classification, Classification method, Joint angles, Low-cost devices, Lower body, Lower extremity, Motion Classification, Motion Vector(MV), Robot Control, Rule-based, Smart devices