PERFORMANCE CHARACTERISTICS AND MODELING OF POWER RECEIVED OF GSM SIGNAL: (A CASE STUDY OF BENIN CITY)

ABSTRACT
This work is on performance characteristics and modeling of power received of GSM signal using Benin City as a case study.
The study methodology employed here consists of measurement of power received and its associated location parameters referenced globally. The data was analyzed to determine coverage level for different operators using mean opinion scale. In order to conduct an assessment on the ability of the mobile networks to setup and hold calls (accessibility index) in Benin City, six thousand and sixteen calls were initiated in forty-seven locations throughout the City between January 2004 and March 2005. Seventy five percent of the calls were inter-network, while the remaining 25% were intra network. The propagation path loss characteristics of Benin City were investigated using fifteen different environments which reflect an exhaustive measurement and good representation of the City. Consequently power received (Pri) was measured from a distance (d) from the base station for various environments investigated. The data was analyzed to determine the propagation path exponent and path loss characteristics.
The use of both empirical and deterministic model was employed in developing a model equation to predict power received in Benin City. The deterministic prediction involved solving a special parabolic equation using finite element method derived from Maxwell equation with backward propagation ignored.
The results of this work reveal that mobile networks operators in Benin City offered a good level of coverage and intra-accessibility for the period under investigation.
However the inter-connectivity assessment reveals that more work needs to be done among the mobile network operators. The propagation path exponent for Benin City was found to be 3.3. The two models developed were compared with both the measured values and existing models. The results of which proved satisfactory. The deterministic model was found to predict fairly better than the empirical one with error of 5.50% and 5.82% respectively when compared with the measured values