Artificial Neural Network Based Indoor Positioning in Visible Light Communication Systems

Cavdar A., TÜRK K.

International Conference on Artificial Intelligence and Data Processing (IDAP), Malatya, Türkiye, 28 - 30 Eylül 2018 identifier

  • Cilt numarası:
  • Basıldığı Şehir: Malatya
  • Basıldığı Ülke: Türkiye


In this study, an artificial neural network (ANN) based visible light positioning system is proposed. For this purpose, an empty and closed room scenario is considered. In this scenario, a LED (light emitting diode) bulb is the transmitter and a photodiode is the receiver. The communication channel is modelled as a visible light channel which take into acoount of multipath reflections. In this regard, Combined Deterministic and Modified Monte Carlo (CDMMC) method is used. Then the proposed system and the RSS-based system in the literature are compared with each other in terms of positioning performances in indoor environment. The RSS-based indoor positioning system have high positioning errors due to the internal reflections into the room. As a result of this study, the ANN-based positioning system has much higher accuracy and near-realistic results than the RSS-based positioning system has. The disturbing effects of multipath reflections on localization has been considerably removed using ANN-based methods.