Aiming at characteristic of quasi-periodic that motion of human body has in daily life

a motion information collection system of lower limbs which is used for collecting acceleration and foot bottom pressure is designed. Acceleration data features are extracted prior to and immediately after gait events to recognize four motion modes including walking

running

stair ascent

stair descent. Experimental result shows that the average recognition accuracy is 94. 79 %

while only extracting simple time-domain features information classification.