Lower-Body Walking Motion Estimation Using Only Two Shank-Mounted Inertial Measurement Units (IMUs)

Human gait measurement is critical for disease diagnosis and rehabilitation. Dedicated monitoring systems such as optical motion capture systems provide precise gait measurements but are constrained to labs and medical facilities. Wearable sensors such as inertial measurement units (IMUs) are promising for monitoring gaits in personal daily life. However, a large number of wearable sensor units are usually required for comprehensive segment motion measurement and this would introduce disturbance and inconvenience to personal daily activities. In this paper, we explore the feasibility of lower-body gait measurement using only two IMUs, and therefore significantly reducing the number of wearable sensors. One IMU is attached to each shank to capture the angle of the shank and displacement of the ankle joint. A whole-step optimization approach is then used to estimate the trunk displacement and lower-body joint angles based on a seven-segment gait model. The proposed method is validated with multi-subject walking data, achieving root mean square errors of 5.70
_~
7.68 degrees in lower-body joint angle estimation and 2.74 cm in step length estimation. Estimation accuracy is similar to systems using four or more IMUs, demonstrating that inherent gait characteristics can be used to build simple wearable systems.