Differentiated time-frequency characteristics based real-time motion decoding for lower extremity rehabilitation exoskeleton robot

Yuanjie Fan,Yuehong Yin
DOI: https://doi.org/10.1007/978-3-642-33515-0_4
2012-01-01
Abstract:Decode the human motion intension precisely in real time is the key problem in coordinated control of the lower extremity exoskeleton. In this research, the relationship between frequency characteristics of sEMG (surface electromyographic) and muscle contraction is established in real time according to the biomechanism of skeletal muscle; DPSE (Differentiated Power Spectrum Estimation) method is applied to extract frequency characteristics from sEMG precisely and quickly; offset compensation is added to prevent noise disturbance during feature extracting of the sEMG with lower SNR (signal-to-noise ratio). Corresponding experiments on knee joint are conducted by prototype exoskeleton robot. EMGBFT (EMG Biofeedback therapy) based on force and haptic is applied as information feedback. Results show the human-machine interface can decode human motion intension and assist or resist movement of the wearer in real-time.
What problem does this paper attempt to address?