Visible light positioning becomes popular recently. However, its performance is degraded by the indoor diffuse optical channel. An artificial neural-network-based visible light positioning algorithm is proposed in this Letter, and a trained neural network is used to achieve positioning with a diffuse channel. Simulations are made to evaluate the proposed positioning algorithm. Results show that the average positioning error is reduced about 13 times, and the positioning time is reduced about two magnitudes. Moreover, the proposed algorithm is robust with a different field-of-view of the receiver and the reflectivity of the wall, which is suitable for various positioning applications.

We develop an improved global reconstruction method for Fourier ptychographic microscopy, a newly reported technique for wide-field and high-resolution microscopic observation. The gradational strategy and graphic processing unit computing are applied to accelerate the conventional global reconstruction method. Both simulations and experiments are carried out to evaluate the performance of our method, and the results show that this method offers a much faster convergence speed and maintains a good reconstruction quality.

Indoor visible light positioning becomes attractive due to the increasing demands of location-based services. This Letter proposes an indoor imaging visible light positioning scheme with a sampled sparse light source, image sensor, and gyro. An indoor positioning cellular with a single reference light source and an off-the-shelf mobile device is demonstrated. Experimental results show that the 3-dimensional positioning error is only several centimeters even with a rotated, rolled, and pitched mobile device. The proposed scheme is convenient and cost effective because the transmitter takes advantage of the existing lighting infrastructure and the receiver is a commercial mobile phone without any extra accessories.