Results from parametric modeling of the indoor channel frequency response for very large number of sweeps over a frequency range of 3 GHz to 10 GHz are presented. The characterization of the frequency-dependence of the ultra wideband mining channel across the considered bandwidth is addressed and the smooth frequency response is evaluated using an Auto Regression (AR) method, which is proved in this study, sufficient to better approximate each recorded frequency response. This study concludes that the number of poles needed to describe the AR process is always less than 2 in the studied area. The appropriated poles of the model are obtained by minimizing the root means squared error (RMSE) between the original frequency response measurement and the predicted one. It is finally found out that the frequency decaying factor (δ) is around 0.5 in the underground area on study.
[1]
Andreas F. Molisch,et al.
Ultrawideband propagation channels-theory, measurement, and modeling
,
2005,
IEEE Transactions on Vehicular Technology.
[2]
G.Y. Delisle,et al.
Coverage Efficiency of Narrow-Band Wave Propagation in Mining Environments
,
2009,
IEEE Transactions on Electromagnetic Compatibility.
[3]
Sofiène Affes,et al.
Neural Networks for Fingerprinting-Based Indoor Localization Using Ultra-Wideband
,
2009,
J. Commun..
[4]
Z. Irahhauten,et al.
An overview of ultra wide band indoor channel measurements and modeling
,
2004,
IEEE Microwave and Wireless Components Letters.