Classifying gait features for stance and swing using machine learning

Structured gait patterns are currently used as a biometric technique to recognize individuals and in building appropriate exoskeleton technologies. In this study, the features involved in gait were extracted and analyzed. Multiple accelerometers were used to collect the data which was then used to identify gait at various axial positions form healthy volunteers with total of 60 trails. Using machine learning optimal feature sub-selection we analyze data to implicate the optimal methods for analysis of swing phase and stance phase in a closed room environment. Study reports that the accelerometer data could classify based on the accuracy and the efficiency of the learning algorithms. Through feature ranking, results suggest gait can be attributed to a combination of Brachium of arm, Antecubitis, Carpus, Coxal, Femur and Tarsus (Shoulder, Elbow, Wrist, Hip, Knee, and Ankle). This gait study may help analyzing conditions during control and movement-related disease.