Lessons learnt from designing indoor positioning system using 868 MHz radios and neural networks

This paper summarizes our approach and experimental evaluation of infrastructure-based Indoor Positioning System (IPS) designed to be used by First Responders. We are using 868 MHz single channel, power-efficient radio markers and RSSI (Receiver Signal Strength Indicator) fingerprinting. Artificial Neural Network translates vectors of RSSI constructed using mobile units into position. Special preprocessing needs to be applied to on-line signal to construct a vector for classification.