The Push Forward in Rehabilitation: Validation of a Machine Learning Method for Detection of Wheelchair Propulsion Type

Sensors, 2024 · DOI: 10.3390/s24020657 · Published: January 19, 2024

Assistive Technology Bioinformatics Rehabilitation

Simple Explanation

This study focuses on developing a simple method to monitor wheelchair use, distinguishing between active (self-propelled) and passive (attendant-pushed) propulsion. The method uses machine learning to analyze data from sensors placed on the wheelchair to detect the type of movement. The goal is to provide an easy-to-use tool for rehabilitation patients to monitor their activity levels and for clinicians to track patient progress.

Study Duration

Not specified

Participants

24 in-patient MWUs (2 females, 22 males, age of 57.8 ± 12.7)

Evidence Level

Cross-sectional and observational

Key Findings

1
The machine learning method showed high accuracy in detecting the type of wheelchair propulsion, with an F1 score of 0.886 when using both frame and wheel sensors.
2
Even with only a single wheel sensor, the method achieved a good level of accuracy (F1 = 0.827).
3
The study identified key sensor-derived features, such as angular acceleration and velocity around the roll axis, that are important for distinguishing between active and passive wheelchair use.

Research Summary

This study validated a machine learning method for detecting self- or attendant-pushed wheelchair propulsion using data from inertial sensors mounted on the wheelchair. The method achieved high accuracy, especially when using data from both a wheel sensor and a frame sensor, but also performed well with only a single wheel sensor. The validated method can be used to easily determine wheelchair use and the corresponding activity level of patients in rehabilitation, with potential applications in clinical practice and individual monitoring.

Practical Implications

Clinical Application

The developed method can be used in clinical settings to monitor the activity levels of wheelchair users during rehabilitation.

Individual Monitoring

The method can be used as a tool for individual wheelchair users to monitor their activity and promote a more active lifestyle.

Research Tool

The validated method can be used in future research to study activity patterns in wheelchair users and evaluate interventions aimed at increasing physical activity.

Study Limitations

1
All the measurements were performed with the wheelchair provided by the rehabilitation center, so all were of a similar type.
2
Only a small variation in surfaces was included, so no grass, cobblestone, or other pavement.
3
The method still requires some technical support, with data collection in the Movesense Showcase app and post-measurement analysis in a Python script

Your Feedback

Was this summary helpful?

Back to Assistive Technology