A comparison of MLP and RBF neural network architectures for location determination in indoor environments

In this paper two different neural network architectures are investigated for enough accurate position determination of a mobile device in the complex indoor environment. The investigation includes multilayer perceptron (MLP) and radial basis function (RBF) neural networks. It has been already shown for neural networks as powerful tool in RF propagation prediction. The research is based on dependence of the received signal with distance. The neural networks are trained by three training algorithms: scaled conjugate, resilient backpropagation and Levenberg-Marquardit with Bayesian regularization. The obtained results for position prediction show error that is less than 0.25 m.