An intention-based human-exoskeleton EMG interface by using Discrete Wavelet Transform and Principal Component Analysis for sit-to-stand training in gait rehabilitation

Autors: Ebert Choquehuanca, Diego Vasquez, Leonardo Paul Milián Ccopa, David Ronald Achanccaray Diaz.

Presented at: EBMC 2016 – Disney – USA

Abstract.— Assisted robotics and biomedical signal processing are currently two technologies with promissory results worldwide. In that sense, this project proposes to combine both technologies through the development of a human-robot interface based on the acquisition, featuring and pattern recognition of myoelectric signals to control a lower-limb exoskeleton for sit- to-stand training. The aim is to detect the movement intention to set the joint actuators controller by using discrete wavelets packets transform (DWPT) and principal component analysis (PCA) in the developed pattern recognition neural network algorithm. Sit-to-stand movement is the first challenge in gait rehabilitation.

Choquehuanca et al – EMBC 2015 – An intention-based human-exoskeleton EMG interface by using Discrete Wavelet Transform and Principal Component Analysis for sit-to-stand training in gait rehabilitation