DefinePK

Abstract

Predictive propagation models have been extensively employed in the prediction of radio signal coverage, interference analysis and determination of available TV white space for secondary users. Deployment of these predictive models in a different environment other than the one initially designed for has resulted in high prediction errors. Error bounds on the effectiveness of the models at predicting path loss for seven empirical path loss models is investigated for Ondo State in South West Nigeria. The models employed for the analysis are the Hata-Davison, Hata, CCIR, Cost 231, ECC33 and Okumura model and the free space path loss (FSPL) model. Measurements of the signal strengths in four major routes in the Ondo State that span through Urban, Suburban and Rural areas were conducted. The Ondo State Terrestrial TV station on channel 23 operating at 487.25 MHz in the UHF frequency band was employed for this investigation. The measurements were taken along the four routes of concern at some interval and the predicted path loss were then calculated using the five predictive models for the adopted routes and these were compared with the measured path loss. The result was used to calculate and analyse the error bound which was used to determine the mean error, predictive error, spread correlated mean square error, Spearman correlated error and the root mean square error for the adopted predictive models of the routes under consideration. Results from the investigation shows that none of the model gives an RMSE close to the acceptable range of 6-8 dB but the Hata-Davison has the closest value. Therefore, turning or optimizing Hata-Davison model would help in reducing the RMSE values within the acceptable range of 6-8 dB.