Experimental demonstration of an indoor positioning system based on artificial neural network

We propose a 2-D visible light positioning system based on the artificial neural network (ANN), where the light-emitting diodes are grouped into blocks and the block coordinates are encoded with under-sampled modulation. A camera is used to decode the block coordinate in the receiver. The receiver’s position is approximately and precisely estimated using the decoded block coordinate and a typical back propagation ANN, respectively. The experimental results show that the proposed scheme offers a mean positioning error of 1.49 cm.