Kriging based model for path loss prediction in the VHF band

Abstract

Empirical path loss propagation models are widely used for coverage planning and interference analysis of wireless networks. The efficacy of these models depend on the local ambient features of the environments, thus, making them susceptible to high prediction errors when applied in different set of environments. In this paper, we developed a Kriging model for path loss prediction in the Very High Frequency (VHF) band using a geo-statistical approach. Electromagnetic field strength measurements were conducted in the VHF band (i.e. 203.25 MHz) in Ilorin metropolis, Nigeria. The predictions of the proposed model were compared with the already established empirical models. The standard and widely used models that are considered include Hata, COST 231, Egli and ECC-33 models. The performance of each model was adjudged based on the Root Mean Squared Error (RMSE). Results show that across all the distinct routes, the empirical models tend to over-and under estimate path losses while the developed Kriging model recorded the lowest prediction errors. The deployment of Kriging, ECC-33, Hata, COST 231 and Egli models produce an average RMSE of 5.653 dB, 9.351 dB, 14.267 dB, and 18.040 dB dB respectively.