Bluetooth dataset for proximity detection in indoor environments collected with smartphones

Area of Science:

• Computer Science
• Ubiquitous Computing
• Human-Computer Interaction

Background:

• Proximity detection is crucial for context-aware applications in indoor environments.
• Existing datasets may not fully capture the complexities of real-world museum navigation and user interaction.
• Bluetooth Low Energy (BLE) beacons offer a viable technology for indoor localization and proximity sensing.

Purpose of the Study:

• To present a novel dataset for proximity detection research.
• To facilitate the development and evaluation of algorithms for identifying user proximity to points of interest (POI) within an indoor setting.
• To provide a realistic dataset that accounts for variations in user paths and device types.

Main Methods:

• Conducted a data collection experiment involving 32 museum visits.
• Collected Bluetooth beacon messages (Received Signal Strength - RSS) using smartphones.
• Varied smartphone models and visitor paths through 10 distinct artworks.
• Established a detailed ground truth, including start and end times for each artwork visit.

Main Results:

• A comprehensive dataset comprising RSS values, timestamps, artwork identifiers, and ground truth annotations was generated.
• The dataset captures realistic environmental conditions encountered in a museum.
• Enables the study of proximity detection algorithms under varied user behaviors and device characteristics.

Conclusions:

• The released dataset provides a valuable resource for researchers and industry professionals.
• It supports the advancement of algorithms for automatic people and point of interest proximity detection using commercial smartphone technology.
• Accelerates prototyping and development in the field of indoor localization and context-aware systems.