Estimation of the Path-Loss Exponent by Bayesian Filtering Method

Area of Science:

• Wireless Sensor Networks
• Signal Propagation Modeling
• Statistical Signal Processing

Background:

• Accurate parameter estimation is crucial for wireless sensor network (WSN) performance.
• Received Signal Strength Indicator (RSSI) variations pose a significant challenge in signal strength-based systems.
• Log-normal shadowing propagation models are widely used but require precise parameter estimation.

Purpose of the Study:

• To propose a novel algorithm for estimating the path-loss exponent in outdoor WSN environments.
• To evaluate the effectiveness of Bayesian filtering techniques, specifically particle filters, for RSSI data.
• To analyze the stability and accuracy of the proposed dynamic estimation method.

Main Methods:

• Development of a Bayesian filtering algorithm for path-loss exponent estimation.
• Application of particle filters to estimate RSSI data in outdoor experimental settings.
• Comparative analysis of the proposed method against existing techniques and assessment of parameter stability.

Main Results:

• The proposed dynamic estimation algorithm significantly enhances RSSI accuracy compared to experimental measurements.
• The path-loss exponent is shown to be highly dependent on RSSI data.
• Increasing the number of particles in the filter did not proportionally improve the quality of estimated parameters.

Conclusions:

• Bayesian filtering, particularly particle filters, offers a robust approach for estimating path-loss exponents in WSNs.
• The dynamic estimation method provides a substantial improvement in RSSI accuracy for outdoor environments.
• The study highlights the trade-off between computational complexity and accuracy in particle filter applications.