The Long-term Gait Speed Monitoring System Using a Single IMU for a Daily Life Use

Abstract

The gait speed estimation by using inertial measurement unit (IMU) is an efficient method for use of daily life. By attaching one IMU on heel side of the foot, we can easily estimate the user’s walking speed. In order to adapt the estimation to daily life, we have considered long-term gait speed estimation based on a single IMU. Of various types of walking pattern that can be appeared during the long-term daily life walking, this study mainly focuses on the straight line walking because the walking pattern has been widely used to assess the gait function. For accurate estimation of the user’s walking speed, we need to perform a gait cycle segmentation in order to reduce the drift error that comes from double integration of acceleration data of IMU. In this study, the gait cycle is segmented by a firm event, foot-flat. In order to verify the foot-flat detection by using IMU, one subject wore both an IMU and force sensing resistors, and walked on a treadmill. After that four healthy subjects (2M 2F) participated on verifying the gait speed estimation and incline estimation on the treadmill. Finally, four healthy adult subjects (2M 2F) participated on long term gait speed estimation experiment, to verify the excluding algorithm which can exclude various walking patterns such as stair ascent and descent, ramp ascent and descent from long term gait. By extracting the straight line walking on horizontal surface from the long term gait, we can simply acquire the preferred walking speed of the user. The gait speed estimations of the treadmill walking and long term straight line walking were successful. Further studies are needed for accurate estimation and categorization of various gait patterns for the future clinical use. ⓒ 2015 DGIST