Using Bluetooth proximity sensing to determine where office workers spend time at work.

Area of Science:

• Occupational Health
• Wearable Technology
• Human-Computer Interaction

Background:

• Existing wearable movement sensors lack contextual location data in workplaces.
• Accurate location determination is crucial for understanding worker behavior and time allocation.

Purpose of the Study:

• To evaluate the accuracy of Bluetooth sensing using ActiGraph GT9X Link monitors for determining worker location within an office setting.
• To assess the effectiveness of two bespoke algorithms in classifying participant locations based on Bluetooth signals.

Main Methods:

• 30 office workers wore ActiGraph GT9X Link monitors (receivers) and chest cameras for one workday.
• Bluetooth monitors (beacons) were placed in various office locations (entry, kitchen, photocopy room, corridors, office).
• Two algorithms were tested: individual location signal presence and a multi-location heuristic model, validated against video data using F-scores.

Main Results:

• High accuracy (F-score ≈0.97-0.98) was achieved for identifying the office location.
• Accuracy was poor (F-score ≈0.04-0.36) in other locations due to high false positives.
• A multi-location algorithm improved accuracy but remained suboptimal for infrequently visited or high-movement areas.

Conclusions:

• Bluetooth sensing shows potential for identifying primary work locations like offices.
• Multi-beacon placement and heuristic models enhance location classification accuracy.
• Further algorithmic refinement is necessary to improve accuracy in diverse workplace environments.