Sit-to-stand and stand-to-sit transitions (STS), as one of the most demanding functional task in daily life, are affected by aging or stroke and other neurological injuries. Lower-limb exoskeletons can provide extra assistance for affected limbs to recover functional activities [1]. Several studies presented locomotion mode recognition of sitting, standing and STS, or only STS, or static modes [2–6]. They are based on fusing information of the mechanical sensors worn on the human body, e.g. inertial measurement unit (IMU) [2–4], plantar pressure force [5], barometric pressure[2], EMG [6]. However, most of them put sensors on the human body and did not show experiments integrated with exoskeletons. Since the physical interaction between the exoskeleton and human body, the recognition method might be different when wearing a real exoskeleton.

To deal with these problems, in this study we proposed a recognition method about STS based on the multi-sensor fusion information of interior sensors of a light-weight bionic knee exoskeleton (BioKEX). A simple classifier based on Support Vector Machine (SVM) was used considering the computational cost of the processing unit in exoskeleton.

This content is only available via PDF.