WISP, Wearable Inertial Sensor for Online Wheelchair Propulsion Detection

Area of Science:

• Biomechanics
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Manual wheelchair dance is a growing artistic, recreational, and sport activity for individuals with disabilities.
• Propulsion is a key component of wheelchair dance, necessitating methods for monitoring and analysis.
• Current assessment and learning in wheelchair dance lack precise tools for quantifying propulsion quantity and timing.

Purpose of the Study:

• To develop and evaluate a wearable system (WISP) for detecting and characterizing manual wheelchair dance propulsion gestures.
• To enable real-time assessment and provide quantitative feedback for wheelchair dancers.
• To support the teaching and learning process in manual wheelchair dance.

Main Methods:

• A wearable system (WISP) utilizing three inertial sensors placed on the hands and back was developed.
• Two machine learning classifiers were employed for online recognition of basic propulsion gestures (forward, backward, dance).
• A conditional block reconstructed eight specific propulsion gestures, with an online paradigm using a sliding window for real-time analysis.

Main Results:

• The WISP system demonstrated high accuracy in recognizing propulsion gestures.
• A two-sensor configuration achieved 90.28% accuracy in identifying propulsion gestures.
• The system successfully quantified propulsions and measured their timing within a dance choreography.

Conclusions:

• The WISP system effectively detects and characterizes manual wheelchair dance propulsion gestures using inertial sensors.
• The two-sensor configuration offers a highly accurate and potentially more practical solution for real-time applications.
• This technology has significant potential applications in the assessment, learning, and teaching of manual wheelchair dance.