Gait Disorder Detection and Classification Method Using Inertia Measurement Unit for Augmented Feedback Training in Wearable Devices.

Hyeonjong Kim,Ji-Won Kim,Junghyuk Ko

doi:10.3390/s21227676

Hyeonjong Kim, Ji-Won Kim + Show 1 more

Open Access

https://doi.org/10.3390/s21227676

Copy DOI

Abstract

Parkinson’s disease (PD) is a common neurodegenerative disease, one of the symptoms of which is a gait disorder, which decreases gait speed and cadence. Recently, augmented feedback training has been considered to achieve effective physical rehabilitation. Therefore, we have devised a numerical modeling process and algorithm for gait detection and classification (GDC) that actively utilizes augmented feedback training. The numerical model converted each joint angle into a magnitude of acceleration (MoA) and a Z-axis angular velocity (ZAV) parameter. Subsequently, we confirmed the validity of both the GDC numerical modeling and algorithm. As a result, a higher gait detection and classification rate (GDCR) could be observed at a higher gait speed and lower acceleration threshold (AT) and gyroscopic threshold (GT). However, the pattern of the GDCR was ambiguous if the patient was affected by a gait disorder compared to a normal user. To utilize the relationships between the GDCR, AT, GT, and gait speed, we controlled the GDCR by using AT and GT as inputs, which we found to be a reasonable methodology. Moreover, the GDC algorithm could distinguish between normal people and people who suffered from gait disorders. Consequently, the GDC method could be used for rehabilitation and gait evaluation.

Highlights

Parkinson’s disease (PD) is one of the most common neurodegenerative diseases among the elderly
We developed a device to determine whether the current gait could be qualitatively satisfied using changes in the magnitude of acceleration (MoA) and Z-axis angular velocity (ZAV) that occurred near the ankle joint during a gait, and to proceed with augmented feedback training by generating a signal that the gait had been recognized and classified if it were qualitatively satisfied
We developed a numerical model that could convert the joint angles into a MoA and ZAV, and developed an algorithm that could evaluate a gait disorder based on the premise that the pattern of PD patients who suffer from a gait disorder is different from that of normal people

Summary

Introduction

Parkinson’s disease (PD) is one of the most common neurodegenerative diseases among the elderly. Patients suffer from motor symptoms—such as bradykinesia, tremors, freezing of gait, and non-motor symptoms—such as insomnia and depression. These symptoms diminish the quality of life (QOL) of patients and caregivers. Medical staff—such as the doctor in charge—usually prescribe medication and rehabilitation to patients with PD, with rehabilitation being considered a supportive role to medication [4]. Recent research [5,6,7,8] has provided evidence that rehabilitation can relieve motor and non-motor symptoms, improving a patient’s QOL. Rehabilitation has recently been considered a primary and independent PD treatment method—that is not a supportive treatment method—which can prevent neural degradation [9,10]

Methods

Results

Discussion

Conclusion