Toward Convenient and Accurate IMU-based Gait Analysis

While inertial measurement unit (IMU)-based systems have shown their potential in quantifying medically-significant gait parameters, it remains to be demonstrated that they can provide accurate and reliable parameters both across various walking conditions and in healthcare settings. Using an IMU-based system we have previously developed, with one IMU module on each subject’s heel, we quantify gait parameters in 55 men and 46 women, all healthy and aged 40-65, in normal, dual-task, and fast-walking conditions. We evaluate their intra-session reliability, and we establish a new reference database of such parameters showing good to excellent reliability. ICC(2,1) assesses relative reliability, while SEM% and MDC% evaluate absolute reliability. The reliability is excellent for all spatiotemporal gait parameters and stride length (SL) symmetry ratio (ICC≥0.90, SEM%≤4.5%, MDC%≤12.4%) across all conditions. It is good to excellent for fast-walking performance (FWP) indices of stride (Sr), stance (Sa), double-support (DS), and step (St) times; gait speed (GS), and GS normalized to leg length (GSn1) and body height (GSn2) (ICC≥0.91, |SEM%|≤10.0%, |MDC%|≤27.6%). Men have higher swing time (Sw) and SL across all conditions. The following parameters are gender-independent: (1) Sa, DS, GSn1, GSn2; (2) symmetry ratios of Sa, Sa ratio (Sa%), Sw, Sw ratio (Sw%), SL, GS; and (3) FWPs of Sr, Sa, Sw, DS, St, cadence, Sa%, Sw%. Our results provide reference values with new insights into gender FWP comparisons rarely reported in the literature. The advantages and reliability of our IMU-based system make it suitable in medical applications such as prosthetic evaluation, fall-risk assessment, and rehabilitation.