earticle

영어

Simulated gait experiments have provided evidence of the possibility of falls when test subjects experienced abnormal gait (fluctuating gait cycle). Falls will lead to increased healthcare and social cost. This study explored the possibility to classify normal gaits (stable gait cycle) and abnormal gaits. A triaxial accelerometer was used to capture 3-dimensional values of trunk acceleration data for 144 healthy subjects. Normal and abnormal gait experiments were carried out and the experiment data was analyzed statistically. Quantitative analysis results revealed significant differences between the values of trunk acceleration of normal and abnormal gaits. The values of trunk acceleration of abnormal gaits in medio-lateral, anterior-posterior and vertical directions are 257%, 376% and 217% larger than those of a normal gait respectively. A threshold based algorithm to classify normal and abnormal gaits was proposed and evaluated by the developed prototype classifier using the smartphone. The prototype classifier has achieved 100% accuracy in the ability to classify normal and abnormal gaits.