High-Accuracy 3D Indoor Visible Light Positioning Method Based on the Improved Adaptive Cuckoo Search Algorithm

In visible light positioning(VLP) system, in the moving process of photodiode(PD), PD will rotate a small angle to a certain extent, although the rotation angle is small, it will still cause a large positioning error, thus, in order to alleviate the error caused by the rotation of PD, a high-precision 3D indoor VLP method based on the improved adaptive cuckoo search (VLP-IACS) algorithm is proposed in this paper. Firstly, the rotation angles of the photodiode (PD) are introduced into the optical channel transmission model instead of assuming that the PD and the light-emitting diode (LED) are parallel to each other. Secondly, two adaptive strategies are applied to update the detection probability pa\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p_{a}$$\end{document} and step factor α0\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha_{0}$$\end{document} in the traditional cuckoo search (CS) algorithm, and the convergence speed of the cuckoo search algorithm is significantly enhanced. Finally, the IACS algorithm is successfully applied to solve the 3D indoor positioning problem in an indoor space with dimensions of 5 m ×\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times$$\end{document} 5 m ×\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times$$\end{document} 6 m. Simulation results for fixed positioning show that in the case of no PD rotation, the average 3D positioning error is 2.20 cm, and in the case of PD rotation, the average 3D positioning errors under different rotation angle ranges are 9.04 cm, 14.45 cm and 16.22 cm. The results of kinematic positioning show that in the case of no PD rotation, the average 3D positioning error is 1.54 cm, and in the case of PD rotation, the average 3D positioning error is 16.48 cm. The proposed method can effectively reduce the degradation caused by PD rotation in the positioning system and can potentially be used in various indoor positioning scenarios.