Wearable Sensor-Based Real-Time Gait Detection: A Systematic Review

Hari Prasanth, Miroslav Caban, Urs Keller, Grégoire Courtine, Auke Jan Ijspeert, Heike Vallery, Joachim von ZitzewitzSensors 2021 · DOI · OpenReview

TL;DR

This systematic review synthesizes the current landscape of wearable sensors and algorithms for real-time gait detection, a crucial technology for gait rehabilitation and assistive devices. The authors found that Inertial Measurement Units (IMUs) are the dominant sensor, rule-based methods are the most common algorithms, and heel strike and toe off are the most frequently detected gait events.

What Problem Does This Paper Address?

Traditionally, gait analysis required specialized laboratory equipment, limiting its accessibility and application in real-world scenarios. The advent of wearable sensors has opened doors for continuous, real-time gait monitoring outside the lab, which is essential for developing advanced gait rehabilitation techniques and assistive devices like neuroprostheses. However, the field is fragmented, with a lack of clear understanding regarding the most effective sensor-device combinations and algorithmic approaches, especially for individuals with gait impairments. This paper aims to provide a comprehensive overview and systematic analysis of existing research to guide future development and clinical application.

Methods

The authors conducted a systematic literature review across four major scientific databases (Scopus, Web of Science, Cochrane, and PubMed). They identified 1262 articles, from which 113 were analyzed in full-text. The review focused on intra-stride temporal gait features (ISTGFs) as they are crucial for real-time control applications. The analysis categorized studies based on gait events and phases detected, wearable sensors used, and the real-time gait analysis algorithms employed. They also extracted data on studies that validated their methods on impaired populations.

Datasets

Not discussed in the paper. The paper focuses on a systematic review of existing literature rather than introducing new datasets.

Key Insights and Ideas

  • IMUs Reign Supreme: Inertial Measurement Units (IMUs) are overwhelmingly the most popular choice for wearable gait sensing, likely due to their versatility, relatively low cost, and ease of placement.
  • Rule-Based Simplicity Wins: Rule-based methods, often relying on threshold or peak detection, are the most prevalent algorithms. Their simplicity and low computational complexity make them well-suited for real-time applications where latency is critical.
  • Heel Strike & Toe Off are Key: The most commonly sought-after gait events are heel strike and toe off, indicating their importance for understanding gait dynamics.
  • Shank & Foot are Prime Real Estate: For IMU placement, the shank and foot are the most frequently chosen body segments, suggesting they offer rich gait information.
  • Validation Gap for Impaired Gait: While many studies anticipate their methods working for pathological gait, less than a third actually validate them on impaired subjects, highlighting a significant gap for clinical translation.
  • IPS as the Gold Standard (for Validation): Insole Pressure Sensors (IPS) are frequently used as the "gold standard" for ground-truth validation of IMU-based gait detection, making them a crucial component in the research ecosystem.

Key Results

  • Out of 1262 identified articles, 113 were analyzed in full-text, with 99 used for quantitative analysis.
  • Gait Events: Heel strike (45 studies) and toe off (42 studies) were the most frequently detected gait events.
  • Gait Phases: Swing (22 studies) and mid-stance (17 studies) were the most commonly identified gait phases.
  • Sensor Usage: IMUs were used in 77% of studies, either alone or in combination with other sensors. IPSs were used in 57% of studies.
  • Algorithm Usage: Rule-based methods were employed by 63% of studies, followed by machine learning (19%) and other methods.
  • Validation: Only 31 out of 99 studies validated their methods on impaired subjects.
  • Recommended Combination: The authors recommend a combination of IMU sensors and rule-based algorithms as an optimal solution for clinical applications, with IPS used for ground-truth validation.

Limitations and Future Work

  • Quantitative Comparison Difficulty: The heterogeneity of evaluation criteria across studies prevented a direct quantitative comparison of all methods. Future studies are encouraged to report standardized performance metrics.
  • Clinical Application Gap: While the review focused on technical aspects, the authors note that algorithm performance can be heavily influenced by gait impairments, suggesting that future work should more explicitly investigate clinical applications and how impairments dictate algorithm choice.
  • Need for Impaired Gait Validation: The limited validation on impaired subjects is a significant limitation. The authors emphasize the crucial need for studies to evaluate their methods on target populations and across relevant conditions, including the use of walking aids.
  • Usability and Durability: While IMUs are popular, the paper touches upon the usability constraints of other sensors like EMG and the durability concerns of IPS. Future work could focus on developing more user-friendly and robust sensor solutions.

Summary updated

Abstract

Gait analysis has traditionally been carried out in a laboratory environment using expensive equipment, but, recently, reliable, affordable, and wearable sensors have enabled integration into clinical applications as well as use during activities of daily living. Real-time gait analysis is key to the development of gait rehabilitation techniques and assistive devices such as neuroprostheses. This article presents a systematic review of wearable sensors and techniques used in real-time gait analysis, and their application to pathological gait. From four major scientific databases, we identified 1262 articles of which 113 were analyzed in full-text. We found that heel strike and toe off are the most sought-after gait events. Inertial measurement units (IMU) are the most widely used wearable sensors and the shank and foot are the preferred placements. Insole pressure sensors are the most common sensors for ground-truth validation for IMU-based gait detection. Rule-based techniques relying on threshold or peak detection are the most widely used gait detection method. The heterogeneity of evaluation criteria prevented quantitative performance comparison of all methods. Although most studies predicted that the proposed methods would work on pathological gait, less than one third were validated on such data. Clinical applications of gait detection algorithms were considered, and we recommend a combination of IMU and rule-based methods as an optimal solution.
BibTeX
@article{prasanth2021wearable,
  author = {Hari Prasanth and Miroslav Caban and Urs Keller and Grégoire Courtine and Auke Jan Ijspeert and Heike Vallery and Joachim von Zitzewitz},
  title = {{Wearable Sensor-Based Real-Time Gait Detection: A Systematic Review}},
  journal = {Sensors},
  year = {2021},
  doi = {10.3390/s21082727},
}
Figure