An Adaptive Bluetooth/Wi-Fi Fingerprint Positioning Method based on Gaussian Process Regression and Relative Distance

Area of Science:

• Indoor positioning systems
• Wireless sensor networks
• Signal processing

Background:

• Accurate indoor positioning is crucial for various applications.
• Fusion of Bluetooth fingerprint positioning (BFP) and Wi-Fi fingerprint positioning (WFP) requires trusted data.
• Existing methods struggle with integrating heterogeneous positioning data effectively.

Purpose of the Study:

• To propose an adaptive method for fusing BFP and WFP using Gaussian process regression (GPR) and relative distance (RD).
• To develop a system that can select trusted positioning results for fusion, improving overall accuracy.
• To enhance the reliability and precision of indoor positioning systems.

Main Methods:

• Construction of Bluetooth and Wi-Fi fingerprint databases using received signal strength (RSS) measurements.
• Building fingerprint positioning error prediction models with GPR.
• Online determination of trusted positioning results using RD and prediction models.
• Fusion strategy: selecting a single trusted result or the mean of both if equally trusted/untrusted.

Main Results:

• The proposed adaptive method achieved a mean positioning error of 2.06 m.
• A root-mean-square error (RMSE) of 1.449 m was recorded.
• The method demonstrated superior performance compared to standalone BFP and WFP.

Conclusions:

• The adaptive GPR and RD-based method effectively fuses BFP and WFP by selecting trusted positioning results.
• This approach significantly improves indoor positioning accuracy and reliability.
• The proposed method offers a robust solution for integrated wireless positioning systems.