Ankle-motion-related artifact in shank-mounted IMU posture measurements during gait

Shank-mounted inertial measurement units (IMUs) are attractive for portable gait analysis, but their posture estimates may be affected by attachment-related artifact associated with soft tissue and strap motion around the ankle. This study examined whether ankle joint motion contributes to this artifact in a distal shank-mounted configuration. Two independent gait datasets collected using different measurement and attachment configurations were analyzed. DS1 included 15 healthy young adults walking shod and barefoot with 13,563 analyzed steps, and DS2 included 51 healthy adults walking barefoot with 2634 analyzed steps. In both datasets, IMU-based shank posture was compared with optical motion capture (OMC), and the observed IMU-OMC difference was related to ankle angle using correlation, linear regression, and ICC before and after regression-based correction. The IMU-OMC difference was strongly related to ankle angle in DS1-shod (r = 0.71), moderately related in DS1-barefoot (r = 0.54), and weakly related in DS2-barefoot (r = 0.21). Regression coefficients were 0.159, 0.096, and 0.035 for DS1-shod, DS1-bf, and DS2-bf, respectively. Regression-based correction improved ICC from 0.9982 to 0.9995 in DS1-shod and reduced stride-level error variability from 1.84° to 1.27°, whereas effects in DS2 were minimal. These findings indicate that ankle-motion-related artifact can affect posture estimates from distally mounted shank IMUs, but its magnitude varied across datasets. Controlled experiments are needed to identify the responsible configuration factors.