New neural network based mobile location estimation in urban propagation models

Abstract

Location estimation finds its applications in many important decisions in cellular networks. Hand offs, cellular fraud detection and location sensitive billing are some of the examples. Many different techniques are currently in use. This work first gives an overview of conventional location estimation techniques and applications, and a new signal-strength based neural network technique is then presented. A mobile architecture based on a simulated urban environment is used to assess the generalization performance of the feed forward multi-layered perceptron (MLP) neural network.