Construction of Hybrid Dual Radio Frequency RSSI (HDRF-RSSI) Fingerprint Database and Indoor Location Method.

Area of Science:

• Wireless Communication
• Indoor Navigation Systems
• Machine Learning Applications

Background:

• Existing indoor navigation services using radio frequency (RF) struggle to meet public expectations for precision and accuracy.
• Traditional Received Signal Strength Indicator (RSSI) fingerprinting methods often rely on single frequency bands (e.g., 2.4G Wi-Fi).

Purpose of the Study:

• To propose a novel method for constructing a hybrid dual frequency RSSI (HDRF-RSSI) fingerprint library for improved indoor positioning.
• To enhance the accuracy and precision of smartphone indoor positioning by leveraging dual-band Wi-Fi signals.

Main Methods:

• Constructed a HDRF-RSSI fingerprint library by combining 2.4G and 5G RF signals from the same access point (AP), doubling fingerprint dimensions.
• Developed a hybrid RF fingerprinting model, training loss function, and location evaluation algorithm using machine learning.
• Collected dual RF RSSI fingerprint data and trained the model in an experimental scene.

Main Results:

• The HDRF-RSSI fingerprinting demonstrated an 18.1% higher feature discriminability compared to traditional 2.4G RF RSSI fingerprinting.
• The proposed machine learning method effectively improved the precision and accuracy of indoor positioning.
• The method overcomes limitations where transmission points (TP) and APs must be visible for positioning.

Conclusions:

• Combining 2.4G and 5G RF RSSI vectors significantly enhances indoor positioning performance for smartphones.
• The HDRF-RSSI fingerprint library and associated machine learning algorithms offer a more robust solution for indoor navigation.
• This approach provides a practical advancement for achieving higher accuracy in wireless indoor positioning systems.