IKULDAS: An Improved kNN-Based UHF RFID Indoor Localization Algorithm for Directional Radiation Scenario

Area of Science:

• Engineering
• Computer Science

Background:

• Ultra-high frequency radio frequency identification (UHF RFID) is crucial for indoor localization and context-awareness.
• Existing systems often use simplified models and k-nearest neighbor (kNN) algorithms, facing performance degradation due to unconsidered antenna directionality and suboptimal k value selection.

Purpose of the Study:

• To develop an improved kNN-based indoor localization algorithm (IKULDAS) that addresses directional radiation in UHF RFID systems.
• To enhance positioning accuracy and efficiency in context-aware services.

Main Methods:

• Developed a novel RSSI estimation model incorporating dipole and patch antenna gain features.
• Introduced inclination and rotation angles to model antenna postures and re-expressed gains for direct and reflection paths.
• Embedded three strategies into kNN: optimal fixed rotation angle filtering, NJW-based nearest-neighbor tag extraction, and a dynamic mapping mechanism for accelerated tracking.

Main Results:

• The proposed IKULDAS algorithm demonstrates superior positioning accuracy compared to traditional algorithms.
• IKULDAS achieves significantly lower time consumption, indicating improved efficiency.
• The novel RSSI estimation model effectively accounts for directional radiation, mitigating performance degradation.

Conclusions:

• IKULDAS offers a more robust and accurate solution for UHF RFID-based indoor localization, especially in scenarios with directional antennas.
• The algorithm's enhancements in accuracy and speed make it highly suitable for demanding context-aware applications.
• This work provides a valuable advancement in indoor localization technology by addressing key limitations of previous methods.