SDR-Fi-Z: A Wireless Local Area Network-Fingerprinting-Based Indoor Positioning Method for E911 Vertical Accuracy...

Area of Science:

• Wireless Communications
• Indoor Positioning Systems
• Machine Learning Applications

Background:

• The Federal Communications Commission's Enhanced 911 (E911) mandate necessitates improved indoor 3D location accuracy for emergency calls.
• Existing indoor localization systems often rely on Received Signal Strength Indicators (RSSIs) from Wireless Local Area Networks (WLANs), but signal fluctuations limit accuracy, especially for multi-floor or 3D positioning.
• Channel State Information (CSI) offers a more robust alternative to RSSIs for indoor localization due to its sensitivity to the wireless channel's physical characteristics.

Purpose of the Study:

• To investigate the feasibility of using Wi-Fi Channel State Information (CSI) for accurate Z-axis (vertical) indoor positioning.
• To evaluate the performance of machine learning models, specifically Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN), in estimating 3D and vertical positions using CSI data.
• To assess the potential of CSI-based localization in meeting E911 accuracy compliance for emergency services.

Main Methods:

• Collected Channel State Information (CSI) measurements from Wi-Fi signals within an indoor environment.
• Employed two machine learning algorithms: an Artificial Neural Network (ANN) and a Convolutional Neural Network (CNN).
• Trained and tested the models to estimate both three-dimensional (3D) coordinates and vertical (Z-axis) location using the CSI data.

Main Results:

• The study demonstrated that CSI measurements can be effectively utilized for indoor vertical (Z-axis) location estimation, an area less explored than horizontal positioning.
• Both ANN and CNN models showed capability in estimating 3D positions, with a particular focus on the feasibility of achieving E911 vertical accuracy.
• Results indicate that CSI-based localization holds significant potential for enhancing the accuracy of indoor positioning systems for emergency response.

Conclusions:

• Channel State Information (CSI) is a promising signal characteristic for high-accuracy indoor 3D localization, especially for vertical positioning critical for E911 compliance.
• Machine learning techniques, including ANN and CNN, are effective tools for processing CSI data to achieve precise indoor location estimation.
• This research provides a foundation for developing advanced indoor positioning systems that meet the stringent accuracy requirements of modern emergency communication mandates.