Atomic Head Movement Analysis for Wearable Four-Dimensional Task Load Recognition.

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Occupational Health

Background:

• Wearable sensors effectively monitor diverse physical activities for health.
• Sedentary office work, often physically similar, requires specific monitoring methods.
• Head movement analysis offers a novel sensing modality for office environments.

Purpose of the Study:

• To explore head movement as a sensing modality for analyzing physical and mental activity.
• To develop an algorithm for segmenting gyroscope signals into atomic head movement events.
• To recognize four dimensions of task load: cognitive, perceptual, communicative, and physical.

Main Methods:

• Collected head movement data from 24 participants using tri-axial inertial sensors.
• Developed a novel algorithm to segment gyroscope signals into atomic head movement events.
• Classified four dimensions of task load (cognitive, perceptual, communicative, physical) based on head movements.

Main Results:

• Achieved 70% accuracy in recognizing cognitive load levels.
• Exceeded 80% accuracy in recognizing perceptual, communicative, and physical load.
• Proposed head movement event features outperformed 181 features from prior studies.
• Demonstrated diagnostic capability of atomic event features across different load types.

Conclusions:

• Head movement analysis is a promising method for monitoring physical and mental workload in office settings.
• The developed algorithm and features enable accurate assessment of task load dimensions.
• This approach holds potential for improving health monitoring and workload management in occupational environments.