Comparative analysis of the performance of regression machine learning models for indoor visible light positioning systems

• Published in: Discover applied sciences Vol. 7; no. 9; pp. 933 - 24
• Main Author: Moharam, Mohamed Hussien
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2025; Springer Nature B.V; Springer
• Subjects: Accuracy; Algorithms; Applied and Technical Physics; Artificial intelligence; Ceilings; Chemistry/Food Science; Comparative analysis; Decision trees; Distribution functions; Earth Sciences; Effectiveness; Engineering; Environment; Error correction; Gated recurrent unit (GRU); Indoor environments; Learning algorithms; Light emitting diodes; Localization; Long short-term memory (LSTM); Machine learning; Materials Science; Mean absolute percentage error (MAPE); Methods; Neural networks; Optical communication; Optimization; Performance evaluation; Photoelectricity; Real time; Regression analysis; Root-mean-square errors; Visible light positioning (VLP); Mean absolute percentage error (MAPE); Gated recurrent unit (GRU); Visible light positioning (VLP); Long short-term memory (LSTM)
• ISSN: 3004-9261; 2523-3963; 3004-9261; 2523-3971
• DOI: 10.1007/s42452-025-06519-y

This paper presents an Indoor Visible Light Positioning (VLP) system designed for deployment in enclosed environments, using four ceiling-mounted Light Emitting Diodes (LEDs) to serve both illumination and positioning functions. Each LED is fixed at predefined coordinates and transmits unique signals to a floor-level receiver via Visible Light Communication (VLC) technology. Upon receiving these signals, the system performs photoelectric conversion, translating light signals into electrical signals, and generating the dataset used for training machine learning models. Several models, including LSTM, GRU, Random Forest, KNN, Decision Tree, and XGBoost, were trained and evaluated for positioning accuracy. The experimental results indicate that XGBoost achieved the best performance, with remarkably low error rates, producing a MAPE of 0.0022%, an RMSE of 0.0011, and a perfect R 2 score of 1, thus being the most effective model for this application. LSTM and GRU are neural network-based models that performed very close to XGBoost. XGBoost's exceptional error correction capabilities and consistent performance across all evaluation metrics make it particularly suitable for high-precision indoor VLP systems, demonstrating superior ability in handling complex spatial relationships and the inherent variability of indoor environments. The system's effectiveness in utilizing light signals for precise node positioning marks a significant advancement in the field of indoor positioning, offering a reliable solution for real-world applications where high accuracy is essential. Graphical Abstract Article Highlights A novel indoor visible light positioning (VLP) system is developed using four ceiling-mounted LEDs and machine learning models, enabling precise positioning through visible light communication as a cost-effective alternative to conventional radio frequency-based methods. The evaluation of six machine learning models highlights XGBoost as the most effective, with LSTM and GRU also demonstrating strong predictive capabilities, emphasizing their suitability for accurate indoor positioning. Detailed performance analysis, utilizing visualization techniques such as error distributions, cumulative distribution functions, and 3D scatter plots, offers valuable insights into model behavior and confirms the system's reliability for practical applications.