Indoor Positioning System Based on Visible Light Communications

Abstract

In this work is proposed an implementation of a bidimensional indoor positioning system based on visible light communications. The system needs at least three led lamps forming a triangle to localize objects. The trilateration is used to calculate the position of the object, and the RSS to estimate the distance between the object and the lights. The estimators used are deterministic and computational models. The deterministic estimator used is a quadratic regression model by least squares. Computational models used are artificial neural networks: one first option includes an ANN for each lamp and a second one a general ANN for the three lamps system. Best results achieved with this system were about 5 cm of accuracy and 2.5 cm of precision, and this can be intended for the positioning of small targets moving in a room.