State of the Field of waist-mounted sensor algorithm for gait events detection: A scoping review

Unsupervised gait phase estimation with domain-adversarial neural network and adaptive window

The performance of previous machine learning models for gait phase is only satisfactory under limited conditions. First, they produce accurate estimations only when the ground truth of the gait phase (of the target subject) is known. In contrast, when the ground truth of a target subject is not used to train an algorithm, the estimation error noticeably increases. Expensive equipment is required to precisely measure the ground truth of the gait phase. Thus, previous methods have practical shortcoming when they are optimized for individual users. To address this problem, this study introduces an unsupervised domain adaptation technique for estimation without the true gait phase of the target subject. Specifically, a domain-adversarial neural network was modified to perform regression on continuous gait phases. Second, the accuracy of previous models can be degraded by variations in stride time. To address this problem, this study developed an adaptive window method that actively considers changes in stride time. This model considerably reduces estimation errors for walking and running motions. Finally, this study proposed a new method to select the optimal source subject (among several subjects) by defining the similarity between sequential embedding features.

Agreement and consistency of five different clinical gait analysis systems in the assessment of spatiotemporal gait parameters

BACKGROUND

Measuring gait function has become an essential tool in the assessment of mobility in aging populations for both, clinicians and researchers. A variety of systems exist that assess gait parameters such as gait cycle time, gait speed or duration of relative gait phases. Due to different measurement principles such as inertial or pressure sensors, accurate detection of spatiotemporal events may vary between systems.

RESEARCH QUESTION

To compare the absolute agreement and consistency in spatiotemporal gait parameters among five different clinical gait analysis systems using different sensor technologies.

METHODS

We compared two devices using inertial sensors (GaitUp & Mobility Lab), two devices using pressure sensor systems (GAITRite & Zebris) as well as one optical system (OptoGait). Twelve older adults walked at self-selected speed through a walkway integrating all of the above systems. Basic spatiotemporal parameters (gait cycle time, cadence, gait speed and stride length) as well as measures of relative phase (stance phase, swing phase, double stance phase, single limb support) were extracted from all systems. We used Intraclass Correlation Coefficients as measures of agreement and consistency.

RESULTS

High agreement and consistency between all systems was found for basic spatiotemporal parameters, whereas parameters of relative phase showed poorer agreement and consistency. Overground measurement (GAITRite & OptoGait) showed generally higher agreement with each other as compared to inertial sensor-based systems.

SIGNIFICANCE

Our results indicate that accurate detection of both, the heel-strike and toe-off event are crucial for reliable results. Systematic errors in the detection of one or both events may only have a small impact on basic spatiotemporal outcomes as errors remain consistent from step to step. Relative phase parameters on the other hand may be affected to a much larger extent as these differences lead to a systematic increase or reduction of relative phase durations.