Multi Linear Regression Model for Mobile Location Estimation in GSM Network

Background/Objectives: Recently, several researchers are directed in the area of mobile location estimation in GSM network. The major interests in the research include improving the accuracy of location estimation. Methods/Statistical analysis: Therefore, this paper present the first application of Multiple Linear Regression (MLR) analysis for mobile location estimation in a GSM network and without pre-processing or manipulating the Location Dependent Parameter (LDP) - Received Signal Strength Indicator (RSSI). The proposed model was developed and evaluated using Received Signal Strength (RSS) and geographical coordinates obtained from drive tests. Findings: The results show that, 67% of the calls, the positioning error is less than 64 m and 95% of the calls will result in positioning error less than 115 m while the maximum error is 275m for the urban area. Application/Improvements: Results show improved accuracy in location estimation. This model can be adopted for any mobile location application including the emergency call services (E-911) that requires very high accuracy level.