Geodesic Path Model for Indoor Propagation Loss Prediction of Narrowband Channels.

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Indoor path loss models are crucial for network planning and sensing applications like localization.
• Traditional models struggle with accurate path loss prediction in complex environments such as ships.
• Accurate path loss prediction is essential for effective path loss-based localization.

Purpose of the Study:

• To develop a novel geodesic pathfinding-based path loss model for improved indoor signal attenuation prediction.
• To address the limitations of traditional models in environments like ships.
• To provide a tool for automatic extraction of path loss model parameters.

Main Methods:

• Developed a pathfinding-based path loss model using geodesics for approximated signal propagation paths.
• Utilized binary floor maps and transmitter locations as input.
• Created a tool to automatically extract parameters like wall crossings, path distance, and direction changes.
• Validated the model against path loss measurements in ship and office environments using DASH-7 sensor networks at 868 MHz.

Main Results:

• Achieved mean absolute errors as low as 4.79 dB and a standard deviation of 3.63 dB in a ship environment.
• The proposed model demonstrated significantly better performance compared to traditional floor-map-based models in the ship environment.
• Showed more modest improvements in an office environment with a mean absolute error of 6.16 dB and a standard deviation of 4.55 dB.

Conclusions:

• The novel geodesic path loss model offers superior accuracy for signal attenuation prediction, particularly in challenging indoor settings like ships.
• The model's ability to use simple inputs (binary floor map, transmitter locations) and its accompanying tool make it practical for applications.
• This work advances path loss modeling for applications such as localization and joint communication and sensing.