The Research of Locomotion-Mode Recognition Based on Multi-SourceInformation and Extreme Learning Machine

Xianghong CAO,Lei LIU,Peng YANG,Yingmao SONG

DOI: https://doi.org/10.3969/j.issn.1004-1699.2017.08.007

2017-01-01

Abstract:The efficient and accurate locomotion-mode recognition is the basis and the key to the flexible prosthesis control.The key is to recognize locomotion-mode(level-ground walking stairs ascent.stairs descent.upslope.downgrade).In order to overcome the problem that a single information source cannot recognize locomotion-mode,human locomotion-mode multi-source information systems were set up.hip joint angle、and acceleration signals and plantar pressure were used as the major source.Human locomotion-mode were decomposed into four segments using the plantar pressure information.hip joint angle features and acceleration signal features were fused into one feature vector according to the user''s locomotion modes characteristics.The features were reduced(at the cost of information content)using principal component analysis(KPCA).These data were initially used to train Extreme Learning Machine(ELM)models,which classify the patterns as upslope,downgrade,stairs ascent,stairs descent or level-ground walking.The average classification accuracy of level-ground walking.stairs ascent.stairs descent.upslope.downgrade is 96.78％.The average recognition time is 0.52 s,The results showed that the method was superior to SVM and BP method.