Title Motion Intention Recognition for Wearable Power Assist System using Multi-Class SVM and Kinematic Model
Authors Masatoshi Kimura (Toyota Technological Institute)
Michihiro Kawanishi (Toyota Technological Institute)
Tatsuo Narikiyo (Toyota Technological Institute)
Abstract This paper is aimed at describing a framework to implement Multi-Class Support Vector Machine (MCSVM)-based motion intention recognition. To this end, we primarily constructed a wearable exoskeleton robot of lower body which is employed as an experimentation platform to test the MCSVM-based motion intention recognition. Having disclosed prototype development and MCSVM, experimental results of motion intention recognition of standing up and seating are presented. We examined the accuracy of method of motion intention recognition based on MCSVM. We also examined utility of adding information of kinematic model to training data.
PDF [Link]