Analyzing Gait With Minimal Body-Worn Sensing: Combining UWB and IMU

Gait disorders are seeing an increase in incidence due to the aging world population. To improve mobility in daily life, it is necessary to accurately monitor gait parameters using body-worn devices. However, current device setups are either too complex or inaccurate. To that end, we propose to combine ultra-wideband (UWB) and inertial measurement units (IMUs) technology as a portable setup on the feet of the user. An extended Kalman filter (EKF) was implemented to fuse the distance information (from UWB) with integration of acceleration and angular velocity measurements to obtain relative feet positions that do not suffer from integration drift. Our results show that spatio-temporal parameters are estimated with higher accuracy (mean step length rRMSE = 0.037, step width rRMSE = 0.049, FIA RMSE = 4.0° and FPA RMSE = 3.4°) using the combination of UWB and IMU sensors compared to using IMUs only: mean step length rRMSE = 0.198, step width rRMSE = 0.456, FIA RMSE = 4.0° and FPA RMSE = 3.3°.