Optimal Propagation Models for Path-loss Prediction in a Mountainous Environment at 2100MHz

Abstract

The necessity and importance of path-loss determination for the purpose of mobile network planning and optimization requires that an optimal model should be developed for each specific environment. This will ensure proper planning leading to satisfactory service delivery to each environment at all times. In this study, 3G (2100MHz) mobile propagation in a mountainous area is investigated in order to determine a suitable model for optimizing path-loss prediction in urban, suburban and rural environments. Five models (free space, COST-231, Hata, Egli and ECC-33) were analyzed and compared with field measurement data. Their Mean Absolute Percentage Error (MAPE) and closeness (Root Mean Square Error) to the measured data were determined and compared. ECC-33 over-predicts path-loss for all the environments. Free space prediction was found to be impracticable. The RMSE of each of Egli, Hata and COST-231 were used to modify the models, predict new path-loss values and compare with the measured path-loss. The new RMSE values showed that the optimized models improved path-loss predictability by 76.42%, 25.50% and 73.67% for the rural, suburban and urban areas of the mountainous environment respectively.