Abstract
In this study, we aimed to realize a simplified method of gait evaluation using a wearable sensor system intended to be used in motor function rehabilitation or daily exercise for healthcare. This paper focused on gait event analysis by using a single gyroscope attached on the instep. Characteristic points such as a turning value or a zero-crossing of the angular velocity waveform measured with the gyroscope were compared with the actual time points of gait events (heel off, toe off, heel contact and foot flat) detected by aluminum electrode sets for normal speed and slow speed walking of three healthy subjects. By using detection rules that could be applied to the offline processing, more than 90% of the gait events were detected within the difference in detection time of 50ms. For the normal speed walking, more than 90% of the gait events were detected within the difference of 42ms by detection rules both in offline and real time processing. These results showed that the characteristic points of the angular velocity waveform corresponded well to the four gait events, allowing to be detected by a single sensor.
