Uplink Measurements Based Positioning of Mobiles in Cellular Networks

The mobile locating technology is capable of determining the geographical location for any mobile subscriber connected to a wireless network. The most important motivations for mobile locating technology are safety and emergency services. In addition to the emergency services, the location technology proved to enable many other services that generate additional revenue, as well as provide help in the reduction of the network maintenance costs. Today, several methods have been proposed and they differ in the underlying location algorithm, achievable accuracy, implementation requirements and costs associated with the deployment.;The goal of this work is to propose a simple and inexpensive mobile locating method for a cellular network.;The approach combines the Base Station (BS) information as well as parameters resulting from the uplink Mobile Assisted Handoff (MAHO) measurements to estimate the mobile location. The performance of the proposed algorithm is evaluated in two different scenarios. The first scenario is implemented completely within the Matlab simulation environment. The simulation platform is developed to test and determine positioning errors in different wireless propagation environments. The second scenario involves a comparison of the basic approach against measured data collected on a live GSM network. The simulation results and live network testing show that the locating errors are large when compared to the FCC's requirements. However, from the standpoint of value added services and network maintenance the performance of the algorithm is adequate.;Several algorithm implementation strategies were considered and it was discovered that the performance of the algorithm fundamentally depends on one's ability to accurately predict the radio signal path loss attenuation. Even small improvements in the propagation modeling accuracy result in large improvements in the algorithm's accuracy. Out of all proposed implementation schemes, the best results were obtained using the maximum likelihood estimation based on the nearest neighbor interpolation method for path loss prediction. In the best case the achieved accuracies are 583 and 1140 meters for 67% and 95% Circular Error Probability (CERP) respectively.