1.上海第二工业大学智能制造与控制工程学院,上海市 201209
2.上海第二工业大学计算机与信息工程学院,上海市 201209
彭煜(1999—),男,硕士生,主要研究方向为可见光定位技术,智能算法在可见光定位中的应用。Email: 1034906697@qq.com
桂林(1981—),男,副教授,博士,主要研究方向为光纤传感技术,可见光定位与传感技术,智能算法在传感与定位技术中的应用。Email: guilin100@yeah.net
收稿:2026-01-22,
修回:2026-04-03,
录用:2026-04-03,
移动端阅览
彭煜,桂林,王进轲. 基于LED角度偏移模型生成指纹库数据的可见光定位方法[J].光子学报,2026,55(7):0712004
PENG Yu, GUI Lin, WANG Jinke. Visible Light Positioning Method Using the Angular Offset model of LED for Fingerprint Database Generation[J]. Acta Photonica Sinica, 2026, 55(7):0712004
彭煜,桂林,王进轲. 基于LED角度偏移模型生成指纹库数据的可见光定位方法[J].光子学报,2026,55(7):0712004 DOI: 10.3788/gzxb20265507.0712004. CSTR: 32255.14.gzxb20265507.0712004.
PENG Yu, GUI Lin, WANG Jinke. Visible Light Positioning Method Using the Angular Offset model of LED for Fingerprint Database Generation[J]. Acta Photonica Sinica, 2026, 55(7):0712004 DOI: 10.3788/gzxb20265507.0712004. CSTR: 32255.14.gzxb20265507.0712004.
针对LED(light-emitting diode)光源出现角度偏移而导致的定位精度降低问题,将角度偏移信息引入朗伯模型中,提出一种估计该角度偏移的方法,使用包含角度偏移的朗伯模型完成指纹库数据的生成,并结合指纹定位方法完成可见光定位。所提角度偏移估计方法,利用序列二次规划SQP(sequential quadratic programming) 算法解决朗伯模型在室内可见光定位的带约束非线性优化问题,再通过非线性自适应粒子群算法NA-PSO(nonlinear adaptive particle swarm optimization)确定方向角,实现在SQP解集中筛选出最优角度偏移值。为了验证所提方法的普适性,仿真环境设置信噪比为10-50dB,仿真结果表明,本文方法定位误差均值较传统WKNN(weighted k-nearest neighbors)和KNN(k-nearest neighbors)降低57.3%,利用角度修正模型显著抑制了角度扰动对接收信号强度(RSS)的影响;与细粒度指纹库WKNN相比,本文方法无需离线阶段采集角度信息,在40-50dB下实现定位精度提升81.4%,倾斜角与方向角估计误差分别降低86.1%和90.7%,在10-20dB高噪声环境定位误差和角度估计误差与使用71万数据量细粒度指纹库定位误差相当,与RBF神经网络相比定位误差均值降低,本文方法在20dB信噪比下CDF(cumulative distribution function)误差累积在中位相较RBF(radial basis function)提升16.7%,在90%分位相较提升40.8%。因此所提室内可见光角度偏移估计方法能有效减小定位误差,在多数噪声情况下具有鲁棒性。
Objective
2
Indoor visible light positioning (VLP) faces significant accuracy degradation due to unintentional LED angular offsets (e.g., during installation or manufacturing). Traditional solutions, such as fine-grained fingerprint databases or neural networks, require extensive offline data collection of angle information, leading to high deployment costs and limited scalability. This study aims to propose a robust, data-efficient method to estimate and compensate for LED angle offsets without offline angle measurements, thereby enhancing VLP accuracy under diverse noise conditions.
Methods
2
The proposed framework, SQP-NAPSO, integrates, Angle Offset Estimation: SQP (Sequential Quadratic Programming)solves a constrained nonlinear optimization problem derived from the Lambertian model to estimate initial azimuth and tilt angles.NA-PSO (Nonlinear Adaptive Particle Swarm Optimization)resolves SQP’s multi-solution ambiguity by fixing the azimuth direction using the indoor point of maximum RSS. NA-PSO enhances convergence and global search via nonlinear adaptive inertia weights and dynamic learning factors. Positioning: The estimated angles refine the Lambertian model to generate an online fingerprint database without offline angle collection. WKNN (Weighted k-Nearest Neighbors) leverages the refined model for robust localization. Experiments validated the method under 10–50 dB SNR, comparing it against baseline approaches: traditional WKNN/KNN, fine-grained WKNN (with offline angle data), and RBF neural networks.Results and Discussions Angle Estimation: NA-PSO outperformed PSO, Greedy Algorithm, and Simulated Annealing in azimuth search accuracy (mean error: 4.4° vs. 13.4°–36.1°). SQP-NAPSO reduced tilt/azimuth estimation errors by 86.1% and 90.7% versus fine-grained WKNN at 40–50 dB SNR. Positioning Accuracy: Mean localization error decreased by 57.3% compared to traditional WKNN/KNN. At 40–50 dB SNR, accuracy improved by 81.4% over fine-grained WKNN (requiring 714k offline data points vs. 63 RSS samples for SQP-NAPSO). In high noise (10–20 dB), accuracy matched fine-grained WKNN while reducing data dependency. Outperformed RBF neural networks: at 20 dB SNR, median CDF error improved by 16.7% and 90th-percentile error by 40.8%. Robustness: Maintained stability across initial angle offsets (e.g., [45°, 10°], [60°, 5°]) and SNR levels. Angle correction suppressed RSS distortion from attitude perturbations, enhancing spatial continuity in fingerprint maps.
Conclusions
2
The SQP-NAPSO method effectively addresses LED angular offsets in VLP systems by: Eliminating costly offline angle data collection, Leveraging model-based optimization for accurate angle estimation under noise (10–50 dB SNR), Achieving >57% positioning error reduction versus classical methods and >81% versus data-intensive fingerprinting. This approach balances high accuracy, computational efficiency, and deployment practicality, making it suitable for real-world indoor positioning applications. Future work may optimize real-time implementation and multi-LED scenarios.
Maheepala M , Kouzani A Z , Joordens M A . Light-based indoor positioning systems: A review [J]. IEEE Sensors Journal , 2020 , 20 ( 8 ): 3971 - 3995 .
Zhu Z , Yang Y , Chen M , et al . A survey on indoor visible light positioning systems: fundamentals [J]. Applications , and Challenges, 2024 .
Zhuang Y , Hua L , Qi L , et al . A survey of positioning systems using visible LED lights [J]. IEEE Communications Surveys & Tutorials , 2018 , 20 ( 3 ): 1963 - 1988 .
叶子蔚 , 叶会英 , 聂翔宇 , 等 . 基于接收信号强度检测的高精度可见光定位方法 [J]. 中国激光 , 2018 , 45 ( 03 ): 263 - 270 .
Jin L , Jiang Q , Yu S , et al . Lightweight RSS-Based Visible Light Positioning Considering Dust Accumulation [J]. Journal of Lightwave Technology , 2025 .
Guoqing H , Ya L , Bo B , et al . Accuracy Evaluation of Indoor AoA Visible Light Positioning With MLP Regression Model [J]. IEEE Photonics Technology Letters , 2025 .
Du P , Zhang S , Chen C , et al . Demonstration of a low-complexity indoor visible light positioning system using an enhanced TDOA scheme [J]. IEEE Photonics Journal , 2018 , 10 ( 4 ): 1 - 10 .
Sun X , Zou Y , Duan J , et al . The positioning accuracy analysis of AOA-based indoor visible light communication system [C]// 2015 International Conference on Optoelectronics and Microelectronics (ICOM) . IEEE , 2015 : 186 - 190 .
Huang H Q , Yang A Y , Feng L H , et al . Artificial neuralnetwork-based visible light positioning algorithm with a diffuse optical channel [J]. Chinese Optics Letters , 2017 , 15 ( 5 ): 20 - 24 .
Mao X , Jing L , Tong Z , et al . Indoor visible light positioning system driven by deep neural network based on Kalman filtering and clustering optimization [J]. Optics Communications , 2025 : 132090 .
赵黎 , 韩中达 , 张峰 . 基于神经网络的可见光室内立体定位研究 [J]. 中国激光 , 2021 , 48 ( 07 ): 145 - 154 . DOI: 10.3788/CJL202148.0706004 http://dx.doi.org/10.3788/CJL202148.0706004
赵楚韩 , 张洪明 , 宋健 . 基于指纹的室内可见光定位方法 [J]. 中国激光 , 2018 , 45 ( 08 ): 202 - 208 .
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 ): 33 - 42 .
秦岭 , 张崇泰 , 郭瑛 , 等 . 基于 Elman 神经网络的可见光室内 定位算法研究 [J]. 光学学报 , 2022 , 42 ( 5 ): 24 - 31 .
ARMSTRONG J , SEKERCIOGLU Y A , NEILD A . Visible light positioning: A roadmap for international standardization [J]. IEEE Communications Magazine , 2013 , 51 ( 12 ): 68 - 73
Wang L , Guo C , Luo P , et al . Indoor visible light localization algorithm based on received signal strength ratio with multi-directional LED array [C]// 2017 IEEE International Conference on Communications Workshops (ICC Workshops) . IEEE , 2017 : 138 - 143 .
李庆香 , 刘海艳 , 陈艺辉 , 等 . 基于神经网络的多朝向 LED 可见光定位 [J]. 光学学报 , 2024 , 44 ( 22 ): 2206004-2206004-11 .
周炳朋 , 陈光森 , 王鑫 . 面向可见光无线定位的指纹推理模型校准 [J]. 电子学报 , 2024 , 52 ( 06 ): 1925 - 1937 .
Chen , Yang , Ming Gao , and Xiao Song . " Method to design the common aperture multi-band optical system based on the PSO algorithm ." Optics express 29.12 ( 2021 ): 18325 - 18335 .
徐世武 , 吴怡 , 王徐芳 . 基于粒子群优化压缩感知的可见光定位算法 [J]. 中国激光 , 2021 , 48 ( 03 ): 81 - 90 .
Guo X , Shao S , Ansari N , et al . Indoor localization using visible light via fusion of multiple classifiers [J]. IEEE photonics journal , 2017 , 9 ( 6 ): 1 - 16 .
刘公绪 , 李龙 , 路建民 , 等 . 无线信号交汇定位的奇异解问题表征与解决 [J]. 中国科学:信息科学 , 2022 , 52 ( 11 ): 1992 - 2010 .
Sung J Y , Yeh C H , Chow C W , et al . Orthogonal frequency-division multiplexing access (OFDMA) based wireless visible light communication (VLC) system [J]. Optics Communications , 2015 , 355 : 261 - 268 .
苏辰希 , 张艳语 , 李盾 , 等 . 基于空间优化的室内可见光定位方法 [J]. 光学学报 , 2024 , 44 ( 13 ): 78 - 87 .
张慧颖 , 于海越 , 王凯 , 等 . 基于KPCA-K-means++和GA-LMS模型的改进RBF神经网络室内可见光定位 [J]. 光学学报 , 2021 , 41 ( 19 ): 72 - 81 . DOI: 10.3788/AOS202141.1906001 http://dx.doi.org/10.3788/AOS202141.1906001
杨广龙 , 孔勇平 , 钟致民 , 等 . 基于多模指纹匹配的室内定位系统设计与实现 [J]. 计算机工程与设计 , 2013 , 34 ( 5 ): 1896 - 1901 .
宋素真 , 桂林 , 杨晨 , 等 . 可见光通信应用于安全支付有效区域的研究 [J]. 上海第二工业大学学报 , 2020 , 37 ( 03 ): 207 - 212 .
0
浏览量
1
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010602201714号