Sensor-Data Fusion for Multi-Person Indoor Location Estimation.

Area of Science:

• Ubiquitous Computing
• Human-Computer Interaction
• Sensor Networks

Background:

• Indoor localization is crucial for monitoring people's movement and activities.
• Tracking individuals who cannot carry wearable devices, such as in caregiving scenarios, presents unique challenges.
• Existing methods often rely on wearable sensors, limiting applicability.

Purpose of the Study:

• To propose a sensor fusion solution for indoor localization.
• To address scenarios where individuals cannot or will not carry smartphones.
• To evaluate the effectiveness of combining anonymous and eponymous sensors for location estimation.

Main Methods:

• Utilized anonymous Passive Infrared (PIR) motion sensors.
• Employed eponymous wearable sensors (smartphones with Estimote beacons).
• Developed a sensor-fusion approach to integrate data from disparate sources.

Main Results:

• Investigated the performance of anonymous-only, eponymous-only, and combined sensor strategies.
• Analyzed the impact of sensor data synchronization on location estimate quality.
• Examined the trade-off between sensor coverage and localization accuracy.

Conclusions:

• Combining anonymous and eponymous sensors offers a robust solution for indoor localization.
• Synchronization issues can be mitigated without device-level mechanisms.
• Sensor coverage and fusion strategies are key to accurate location estimation in complex environments.