View in EDS

Indoor visible light positioning system based on quick response recognition and transformer–long short-term memory fusion.

Saved in:
Bibliographic Details
Title: Indoor visible light positioning system based on quick response recognition and transformer–long short-term memory fusion.
Authors: Liu, Guoqing, Qin, Ling, Hu, Xiaoli, Du, Yongxing
Source: Optical Engineering; Jan2025, Vol. 64 Issue 1, p18102-18102, 1p
Subject Terms: LONG short-term memory, ARTIFICIAL intelligence, VISIBLE spectra, TWO-dimensional bar codes, SHORT-term memory, DEEP learning, LED displays
Abstract: With the continuous development of artificial intelligence technology, visible light positioning (VLP) based on deep learning has become a research hotspot for indoor positioning technology to solve the problems of high complexity, low positioning accuracy, and poor real-time performance of the current positioning system. We propose an indoor visible light localization system based on quick response (QR) code recognition technology and transformer–long short-term memory (LSTM) fusion neural network. The system effectively reduces the system complexity by attaching QR code tags to rectangular light-emitting diode (LED) lamp housings for LED-ID data transmission. During the localization process, the vertex coordinates of the LED projection quadrilateral in the image, the diagonal angles, the distance between the centroid and the image center, and the demodulation results are input into a transformer–LSTM fusion neural network for training. The trained model is then used to predict the coordinates of the test set. The experimental results show that in the experimental area of the 9 m×2.2 m×3 m hospital corridor, the average localization error is 8.75 cm, and the localization time of a single sample point is 4.38 ms, which satisfied the localization accuracy and real-time demand of the indoor environment. [ABSTRACT FROM AUTHOR]
Copyright of Optical Engineering is the property of SPIE - International Society of Optical Engineering and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Complementary Index
Description
Abstract:With the continuous development of artificial intelligence technology, visible light positioning (VLP) based on deep learning has become a research hotspot for indoor positioning technology to solve the problems of high complexity, low positioning accuracy, and poor real-time performance of the current positioning system. We propose an indoor visible light localization system based on quick response (QR) code recognition technology and transformer–long short-term memory (LSTM) fusion neural network. The system effectively reduces the system complexity by attaching QR code tags to rectangular light-emitting diode (LED) lamp housings for LED-ID data transmission. During the localization process, the vertex coordinates of the LED projection quadrilateral in the image, the diagonal angles, the distance between the centroid and the image center, and the demodulation results are input into a transformer–LSTM fusion neural network for training. The trained model is then used to predict the coordinates of the test set. The experimental results show that in the experimental area of the 9 m×2.2 m×3 m hospital corridor, the average localization error is 8.75 cm, and the localization time of a single sample point is 4.38 ms, which satisfied the localization accuracy and real-time demand of the indoor environment. [ABSTRACT FROM AUTHOR]
ISSN:00913286
DOI:10.1117/1.OE.64.1.018102