A Low Complexity Indoor Visible Light Positioning Method

The continuous positioning of a mobile user equipment (MU) in a prompt way is a fundamental requirement in several indoor applications related to ubiquitous computation, assisted living environments, and security/surveillance systems. Motivated by this fact, the current contribution reports a low-complexity visible light positioning (VLP) method suitable for indoor environments in the Big Data era. The proposed architecture consists of multiple light-emitting diodes (LEDs) as light sources and an MU equipped with a photodiode (PD). To guarantee higher spectral efficiency, the LEDs emit sinusoidal waveforms at slightly different predetermined frequencies. The light intensity received at the PD from every LED is continuously estimated after applying a short-time Fourier transformation. To this end, a grid-based numerical technique is proposed to recover the unknown MU position. Additionally, a closed-form solution is presented for the particular case of three LEDs that overrides the need for training points in the grid arrangement. Further, a method for handling noisy signals is proposed, based on averaging the calculated positions from densely overlapping received signals. Finally, a Kalman filter is employed as a post-processing precision improving tool. The proposed VLP efficiency is quantified through respective Monte Carlo simulations that allow the setting of different LED frequencies, received signal processing parameters, MU speed, and noise levels. The results reveal that the suggested approach is robust with significant tolerance in high ambient light levels, computationally efficient, and exhibits low positional error. Finally, to evaluate the performance improvements the new method introduces, we make comparisons against widely-used fingerprint approaches.