A neural network approach for Radio Frequency based indoors localization

Radio Frequency (RF) based localization is appropriate for indoors quasi-structured environments. However some accuracy issues remain which are raised by the characteristics of localization accuracy requirements. This paper adopts the Artificial Neural Network (ANN) method to overcome a few of them. Making use of available communications subsystem built in wireless protocols, Received Signal Strength Indication (RSSI) can become a measured variable that supplies ANN schemes. RSSI post-processing filters improve localization accuracy and ANNs emulate the required trilateration for indoors RF localization. Preliminary ANN learning results are included in this paper. This promising result encourages further research on ANN learning for indoor localization.