Multi-Residential Activity Labelling in Smart Homes with Wearable Tags Using BLE Technology

Area of Science:

• Ubiquitous Computing
• Human-Computer Interaction
• Sensor Networks

Background:

• Smart home platforms offer improved quality of life but face challenges in multi-residential activity labeling.
• Current activity sensors lack individual identification, complicating data labeling in shared living spaces.
• Individual localization is proposed as a solution to automatically label sensor events.

Purpose of the Study:

• To address the challenge of activity labeling in multi-residential smart homes.
• To investigate the impact of Bluetooth low energy (BLE) settings on localization accuracy.
• To propose an automatic activity labeling method integrating BLE localization and ambient sensor data.

Main Methods:

• Utilizing Bluetooth low energy (BLE) tags worn by individuals for broadcasting unique IDs.
• Employing fixed scanners to detect BLE tag broadcasts for individual localization.
• Studying the effects of varying BLE advertising time intervals and power levels on localization accuracy.
• Developing an algorithm for optimal BLE sensor setting selection.
• Integrating BLE localization data with ambient sensor data for automatic activity labeling.

Main Results:

• Demonstrated that BLE localization accuracy is significantly influenced by broadcasting signal strength and time intervals.
• Proposed an efficient algorithm for selecting optimal BLE sensor settings.
• Successfully implemented an automatic activity labeling method in a real multi-resident smart home.
• Validated the effectiveness and applicability of the proposed system.

Conclusions:

• BLE technology is a viable and low-energy solution for individual localization in smart homes.
• Optimized BLE settings are crucial for achieving high localization accuracy.
• The proposed integrated system effectively enables automatic activity labeling in complex smart home environments.