PATH-LOSS PREDICTION MODELS FOR CELLULAR NETWORKS IN OKENE, KOGI STATE, NIGERIA

Abstract

Signal path-loss in Radio communication results in low signal power at the receiver antenna of telecommunication systems. It drastically affects network performance negatively. Path-loss is found to be very high in mountainous environments such as Okene in Kogi State, Nigeria, which is the case study of this research. The result is poor network service despite the fair distribution of network infrastructure. A suitable and accurate path-loss model to predict the path-loss of this environment is necessary to enable network operators make plans for network optimization by deciding the appropriate capacity of equipment, suitable base transceiver station location and efficient network coverage design. The method of modeling in this research is the modification of selected path-loss models which have history of adaptability to various environments and frequencies. The models selected are the Free Space model, the COST-231 model, the Hata model, the Egli model and the ECC-33 model. The predictions by these models were compared with the measured path-loss in three environments – urban Okene central area, suburban Okengwe area and Rural Agasa/Upogoro area. The networks used are 3G UMTS and 2G EDGE of 9mobile, MTN, Glo and Airtel. The average root mean square errors (RMSE) of the predicted path-loss were used to modify the selected models. New path-loss predictions were made using the modified models. The modified models that gave the least average RMSE were selected as the optimal models for that environment at that frequency. The result of the study showed that the optimal models for the urban and suburban environments were developed from COST-231 model for 2G and 3G while optimal models for the rural environment were developed from the Hata model. The average RMSE of the optimal models were 9.39dBm for 3G and 12.88dBm for 2G Networks in the urban area, 11.28dBm for 3G and 11.60dBm for 2G Networks in the suburban area and 10.34dBm for 2G Network in the rural area. However, for 3G in the rural environment, the original Hata model maintained the least average RMSE of 10.02dBm after the modifications process. Hence, no new model was developed for 3G network in the rural area. The optimized models showed various percentages of improvement over the reference models. The range of deviations acceptable for suitability of a model is 0 to 12dBm. The signal power received from the field measurement confirmed that the high signal path-loss was caused by the mountainous terrain. These optimized models are recommended for network operators as a guide to making decisions on network optimization in Okene mountainous area and for planning to upgrade to 4G.