The Use of Bayesian Learning of Neural Networks for Mobile User Position Prediction

Mobility management plays a central role in providing ubiquitous communications services in future wireless mobile networks. In mobility management, there are two key operations, location update and paging, commonly used in tracking mobile users on the move. Location update is to inform the network about a mobile user´s current location, while paging is used for the network to locate a mobile user. Both operations will incur signaling traffic in the resource limited wireless networks. The more frequent the location updates, the less paging in locating a mobile user; thus, there is a trade off in terms of signaling cost In this paper, we present a novel hybrid Bayesian neural network model for predicting locations on cellular networks. We investigate different parallel implementation techniques on mobile devices of the proposed approach and compare it to many standard neural network techniques such as: back-propagation, Elman, Resilient, Levenberg-Marqudat, and one-step secant models. In our experiments, we compare results of the proposed Bayesian neural network with 5 standard neural network techniques in predicting next location. Bayesian learning for neural networks predicts location better than standard neural network techniques since it uses well founded probability model to represent uncertainty about the relationship being learned. The result of Bayesian training is a posterior distribution over network weights.