Efficient Recognition of Informative Measurement in the RF-Based Device-Free Localization

Area of Science:

• Wireless Sensing
• Radio Frequency (RF) Technology
• Localization Systems

Background:

• Device-Free Localization (DFL) uses Radio Frequency (RF) sensing for positioning.
• Back-projection Radio Tomographic Imaging (BRTI) is a lightweight method for real-time DFL.
• Multipath noise in RF networks degrades BRTI performance and requires identifying target-affected links.

Purpose of the Study:

• To enhance RF link state analysis for improved DFL.
• To develop an efficient method for recognizing informative RF links affected by a target.
• To overcome limitations of existing RF link analysis methods in terms of complexity and time.

Main Methods:

• Transformed RF link state analysis into a matrix decomposition problem.
• Proposed an efficient RF link recognition method based on low-rank and sparse decomposition.
• Sensed spatiotemporal variations of RF link states to identify target-affected links.

Main Results:

• The proposed method effectively recognizes RF links reflecting target-induced RSS variations.
• Achieved improved recognition with reduced time complexity compared to existing methods.
• Experimental results demonstrate the method's effectiveness in real-world DFL applications.

Conclusions:

• The low-rank and sparse decomposition method enhances RF link state analysis for DFL.
• Accurate recognition of informative RF links improves BRTI accuracy.
• The proposed method increases the overall efficiency of device-free localization systems.