Indoor positioning in complex environments using modified probabilistic neural network

This paper presents a modified probabilistic neural network (MPNN) based indoor positioning technique, which can be used in complex environment. Firstly, the received signal strengths (RSS) are measured between an object and stations. An average filter is applied to remove noise of RSS set. The extracted RSS features are transformed into reliable distances. Then, A MPNN engine determines coordinate of the object with the input distances. The experiments perform significantly better than triangulation technique when the RSS data are unstable in complex environments.