Indoor localization based on feed-forward Neural Networks and CIR fingerprinting techniques

Many solutions to accurate localization in indoor environments are based on the combination of Channel Impulse Response (CIR), as fingerprinting, and Neural Networks (NN) technique. Suitable there is no theory to indicate the neural network structure for a given problem. It must be done only experimentally. Then, this paper aims to identify the neural network structure and the accompanied algorithm which are suitable for indoor localization. For this purpose, different NN structures accompanied with various algorithms are compared and discussed in order to identify the most appropriate.