Indoor Positioning of Visible Light Communication based on Transformer Model

YAO Yao; ZHU Yali; SHAO Xinyu

doi:10.13756/j.gtxyj.2026.240201

您当前的位置：

首页 >

文章列表页 >

Indoor Positioning of Visible Light Communication based on Transformer Model

更新时间：2026-03-03

- Indoor Positioning of Visible Light Communication based on Transformer Model
- Study on Optical Communications Issue 1, (2026)
- 作者机构：
  
  江苏商贸职业学院　电子与信息学院，江苏　南通　 226011
- 作者简介：
- 基金信息：
- DOI：10.13756/j.gtxyj.2026.240201
  CLC： TP394.1;TH691.9
- Received：13 September 2024，
  
  Revised：2024-10-08，
  
  Published：10 February 2026
- 稿件说明：
移动端阅览
姚瑶，朱亚丽，邵鑫玉. 基于Transformer模型的可见光通信室内定位[J]. 光通信研究，2026(1): 240201.

Yao Y, Zhu Y L, Shao X Y. Indoor Positioning of Visible Light Communication based on Transformer Model[J]. Study on Optical Communications, 2026(1): 240201.
姚瑶，朱亚丽，邵鑫玉. 基于Transformer模型的可见光通信室内定位[J]. 光通信研究，2026(1): 240201. DOI： 10.13756/j.gtxyj.2026.240201.

Yao Y, Zhu Y L, Shao X Y. Indoor Positioning of Visible Light Communication based on Transformer Model[J]. Study on Optical Communications, 2026(1): 240201. DOI： 10.13756/j.gtxyj.2026.240201.

摘要

目的

可见光在室内进行了多径传播时，符号间干扰（ISI）导致基于接收信号强度的室内可见光定位精度不足。针对此问题，文章结合Transformer模型提出了一种新的基于接收信号强度的室内可见光定位系统。

方法

首先，文章推导了室内可见光接收信号强度的数学模型，分析了ISI和室内噪声对室内可见光通信系统的影响；然后，收集了各参考点的接收信号强度和真实坐标构建室内指纹数据库，文章所提神经网络模型在指纹数据库上进行端到端训练，学习从接收信号强度指纹到对应位置坐标的映射关系。

结果

仿真结果表明，文章所提定位系统的定位误差达到厘米级，在地面、0.5、1.0、1.5和2.0 m高度接收面的平均定位误差分别为0.73、0.94、1.13、1.92和2.97 cm，在室内不同高度的定位精度均优于其他基于接收信号强度的定位方法。此外，文章所提系统的平均定位时间为2.3 ms，虽然其速度不具备优势，但所提定位系统的定位精度更高，因此综合性能更具优势。

结论

Transformer模型提取了接收信号强度指纹内部各元素间的依赖性，而且提取了接收信号强度指纹间的依赖性，可挖掘出接收信号强度指纹结构中隐含的空间信息，从而提升了特征关于室内位置的判别性。因此，当接收面高度升至1.5和2.0 m时，文章所提定位系统的平均定位误差并未出现大幅提高的现象。经比较，文章所提系统的累积分布函数图显著好于对比方法。

Abstract

Objective

In the multipath propagation scenario of indoor visible light communication

Inter-Symbol Interference (ISI) leads to low precision of indoor visible light positioning due to low received signal strength. In view of this

a new received signal strength based indoor visible light positioning system combined with Transformer model is proposed.

Methods

Firstly

the indoor visible light received signal strength is analyzed by mathematical derivation

and the impact of ISI and indoor noise on indoor visible light communication systems is analyzed. Then

the received signal strength and real coordinate of each reference point are collected to construct the indoor fingerprint database. The proposed neural network model is end-to-end trained on the fingerprint database

so as to learn the mapping relationship from the received signal strength fingerprint to the corresponding position coordinate.

Results

Simulation results demonstrate that the positioning error of the proposed positioning system reaches the centimeter level

with average positioning errors of 0.73

0.94

1.13

1.92 and 2.97 cm recorded for the receiving planes at ground level

0.5

1.0

1.5 and 2.0 m

respectively. Compared with the other received signal strength based positioning methods

the indoor positioning precision of the proposed system is better at different heights. Besides

the average positioning time of the proposed system is 2.3 ms.

Conclusion

The Transformer model extracts the dependency among the elements of every received signal strength fingerprint. It also extracts the dependency among the received signal strength fingerprints. The hidden spatial information within the received signal strength fingerprint structure is retrievable

effectively boosting the feature discriminability relevant to indoor locations. Hence

the average positioning error of the proposed system does not increase sharply even when the receiving plane height reaches 1.5 and 2.0 m. Moreover

the system's cumulative distribution function curve shows a clear advantage over those of the comparative approaches.

关键词

Keywords

references

邢莉娟 , 李卓 , 王卓 . 自由空间光通信中极化码的应用研究 [J ] . 光通信研究 , 2024 ( 3 ): 230058 .

Xing L J , Li Z , Wang Z . Research on Application of Polar Codes in Free Space Optical Communication [J ] . Study on Optical Communications , 2024 ( 3 ): 230058 .

张慧颖 , 王凯 , 于海越 , 等 . 基于自适应Levy飞行的黄金正弦可见光定位研究 [J ] . 激光技术 , 2022 , 46 ( 4 ): 519 .

Zhang H Y , Wang K , Yu H Y , et al . Research on Golden Sine Visible Light Positioning based on Adaptive Levy Flight [J ] . Laser Technology , 2022 , 46 ( 4 ): 519 .

张慧颖 , 梁誉 , 卢宇希 , 等 . 稀疏训练指纹库融合MMPSO-ELM室内可见光定位 [J ] . 激光技术 , 2022 , 46 ( 6 ): 788 - 795 .

Zhang H Y , Liang Y , Lu Y X , et al . Indoor Visible Light Positioning Using MMPSO-ELM Neural Network based on Sparse Training Fingerprint Database [J ] . Laser Technology , 2022 , 46 ( 6 ): 788 - 795 .

do Nascimento M R F , Coutinho O G G , Olivi L R , et al . A Visible Light Positioning Technique based on Artificial Neural Network [J ] . Journal of Control, Automation and Electrical Systems , 2024 , 35 ( 4 ): 677 - 687 .

赵霞 , 张君毅 , 龙倩倩 . 基于Circle混沌映射的ISSA-ELM神经网络室内可见光定位方法 [J ] . 光学学报 , 2023 , 43 ( 2 ): 0206004 .

Zhao X , Zhang J Y , Long Q Q . Indoor Visible Light Positioning Method Using ISSA-ELM Neural Network based on Circle Chaotic Mapping [J ] . Acta Optica Sinica , 2023 , 43 ( 2 ): 0206004 .

王童瑶 , 肖纯贤 , 林作俊 , 等 . 基于标记LED的单灯室内定位系统设计 [J ] . 光通信研究 , 2022 ( 5 ): 38 - 42 .

Wang T Y , Xiao C X , Lin Z J , et al . The Design of Single LED Indoor Positioning System based on Marked LED [J ] . Study on Optical Communications , 2022 ( 5 ): 38 - 42 .

秦岭 , 刘哲 , 王凤英 , 等 . 基于BPNN-MLR的室内可见光定位算法 [J ] . 光通信研究 , 2022 ( 2 ): 7 - 11 .

Qin L , Liu Z , Wang F Y , et al . Indoor Visible Light Positioning Algorithm based on BPNN-MLR [J ] . Study on Optical Communications , 2022 ( 2 ): 7 - 11 .

Komine T , Nakagawa M . Fundamental Analysis for Visible-Light Communication System Using LED Lights [J ] . IEEE Transactions on Consumer Electronics , 2004 , 50 ( 1 ): 100 - 107 .

刘军 , 王慧民 , 张兴忠 , 等 . 基于Transformer的端到端路面裂缝检测方法 [J ] . 太原理工大学学报 , 2022 , 53 ( 6 ): 1143 - 1151 .

Liu J , Wang H M , Zhang X Z , et al . End-to-End Pavement Crack Detection Method based on Transformer [J ] . Journal of Taiyuan University of Technology , 2022 , 53 ( 6 ): 1143 - 1151 .

Hsu L S , Tsai D C , Chen H M , et al . Using Received-Signal-Strength (RSS) Pre-Processing and Convolutional Neural Network (CNN) to Enhance Position Accuracy in Visible Light Positioning (VLP) [C ] // 2022 Optical Fiber Communications Conference and Exhibition (OFC) . San Diego, CA, USA : IEEE , 2022 : 9748179 .

Chen H , Han W , Wang J , et al . High Accuracy Indoor Visible Light Positioning Using a Long Short Term Memory-Fully Connected Network based Algorithm [J ] . Optics Express , 2021 , 29 ( 25 ): 41109 - 41120 .

Views

下载量

CSCD

CNKI被引量

Alert me when the article has been cited

提交

Tools

Publicity Resources

Indoor Positioning System based on SSA-ELM Neural Network for Visible Light

Non-line-of-sight Visible Light Positioning System based on Deep Learning

Baseline Correction Method of FBG Sensor Network Spectrum based on the Improved LSTM Model

Deep Learning based Channel Estimation in PLC Communication

Application of Deep Learning in Impairments Compensation of Optical Fiber Communication

Related Author

CAO Minghua

PENG Duo

ZHANG Zhicong

YU Kai

NIU Zhen

JIA Kejun

LIN Bangjiang

HUANG Weijie

Related Institution

College of Computer and Communication, Lanzhou University of Technology

Quanzhou Institute of Equipment Manufacturing, Haixi Institutes, Chinese Academy of Sciences

College of Mechanical and Electrical Engineering, Fujian Agriculture and Forestry University

Tianjin University

Beijing Laboratory of Optical Fiber Sensing and System, Beijing Information Science & Technology University

Address：3rd floor, Building 4, Fenghuo Science Park, Gaoxinsilu, East Lake New Technology Development Zone, Wuhan, Hubei Province, P.R.China Postal code：430205
Technical support is provided by Beijing Founder electronics co., LTD 鄂ICP备2022010197 鄂公网安备 42011202002092号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
Total Visits：312958 Visits Today：124

⁰