Biomechanical Gait Variable Estimation Using Wearable Sensors after Unilateral Total Knee Arthroplasty

Abstract

Total knee arthroplasty is a common surgical treatment for end-stage osteoarthritis of the knee. The majority of existing studies that have explored the relationship between recovery and gait biomechanics have been conducted in laboratory settings. However, seamless gait parameter monitoring in real-world conditions may provide a better understanding of recovery post-surgery. The purpose of this study was to estimate kinematic and kinetic gait variables using two ankle-worn wearable sensors in individuals after unilateral total knee arthroplasty. Eighteen subjects at least six months post-unilateral total knee arthroplasty participated in this study. Four biomechanical gait variables were measured using an instrumented split-belt treadmill and motion capture systems. Concurrently, eleven inertial gait variables were extracted from two ankle-worn accelerometers. Subsets of the inertial gait variables for each biomechanical gait variable estimation were statistically selected. Then, hierarchical regressions were created to determine the directional contributions of the inertial gait variables for biomechanical gait variable estimations. Selected inertial gait variables significantly predicted trial-averaged biomechanical gait variables. Moreover, strong directionally-aligned relationships were observed. Wearable-based gait monitoring of multiple and sequential kinetic gait variables in daily life could provide a more accurate understanding of the relationships between movement patterns and recovery from total knee arthroplasty.

Keywords: biomechanical gait variable estimation; inertial gait variable; total knee arthroplasty; wearable sensors.

Figure 1

Framework for developing proposed biomechanical measure estimation models.

Figure 2

Wearable sensor orientation on both legs.

Figure 3

Kinetic biomechanical gait variables: weight-normalized. (a) Knee flexion moment (KFM); (b) knee adduction moment (KAM); (c) anterior/posterior ground reaction forces (aGRF); (d) vertical ground reaction force (vGRF).

Figure 4

Step detection and validation with recognized heel-strikes. (a) Step detection using acceleration in the anterior/posterior direction; (b) validation of step recognition using the ground reaction force in the vertical direction. Note that the intervals between recognized heel-strikes indicate stride cycles.