QA-kNN: Indoor Localization Based on Quartile Analysis and the kNN Classifier for Wireless Networks

Area of Science:

• Wireless Communication
• Localization Algorithms
• Data Pre-processing

Background:

• Received Signal Strength Indicator (RSSI) variation significantly compromises indoor localization accuracy in wireless networks.
• Existing localization methods struggle to mitigate the impact of dynamic environmental factors on RSSI.
• Accurate indoor positioning is crucial for various applications, including asset tracking and navigation.

Purpose of the Study:

• To investigate and propose an improved indoor localization method addressing RSSI variation.
• To enhance the accuracy of indoor localization systems.
• To evaluate the proposed method's performance in both real and simulated environments.

Main Methods:

• Utilized quartile analysis for robust data pre-processing of RSSI measurements.
• Employed the k-nearest neighbors (kNN) classifier for the core localization task.
• Conducted extensive tests in a real-world environment and performed simulations with varied parameters.

Main Results:

• In a real environment with high reference point density (1.284 RPs/m²), the method achieved zero-mean error at test points coinciding with reference points.
• In a simulated environment with lower reference point density (0.327 RPs/m²), a mean localization error of 0.490 m was recorded for random test points.
• The proposed method demonstrated resilience to RSSI variations, maintaining reliable localization performance.

Conclusions:

• The developed indoor localization method effectively mitigates RSSI variation issues.
• The approach shows significant promise for accurate object location in indoor settings.
• The findings contribute valuable insights into enhancing the precision of wireless indoor positioning systems.