Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Transfer Learning for VLC-based indoor Localization: Addressing Environmental Variability

Masood Jan, Wafa Njima, Xun Zhang, Alexander Artemenko

TL;DR

This work tackles the challenge of indoor VLC-based localization under environmental variability in industrial settings. It introduces a transfer-learning framework where a DNN trained on real BOSCH factory data is fine-tuned to adapt to changing lighting conditions, using a combined source/target loss and noise-augmented target data. The approach yields up to 47% accuracy improvements, around 32% energy savings, and roughly 40% faster computation compared with conventional models, while requiring as little as 30% of the data for adaptation. The results demonstrate a scalable, cost-efficient solution for Industry 4.0 environments with dynamic illumination and obstacles.

Abstract

Accurate indoor localization is crucial in industrial environments. Visible Light Communication (VLC) has emerged as a promising solution, offering high accuracy, energy efficiency, and minimal electromagnetic interference. However, VLC-based indoor localization faces challenges due to environmental variability, such as lighting fluctuations and obstacles. To address these challenges, we propose a Transfer Learning (TL)-based approach for VLC-based indoor localization. Using real-world data collected at a BOSCH factory, the TL framework integrates a deep neural network (DNN) to improve localization accuracy by 47\%, reduce energy consumption by 32\%, and decrease computational time by 40\% compared to the conventional models. The proposed solution is highly adaptable under varying environmental conditions and achieves similar accuracy with only 30\% of the dataset, making it a cost-efficient and scalable option for industrial applications in Industry 4.0.

Transfer Learning for VLC-based indoor Localization: Addressing Environmental Variability

TL;DR

This work tackles the challenge of indoor VLC-based localization under environmental variability in industrial settings. It introduces a transfer-learning framework where a DNN trained on real BOSCH factory data is fine-tuned to adapt to changing lighting conditions, using a combined source/target loss and noise-augmented target data. The approach yields up to 47% accuracy improvements, around 32% energy savings, and roughly 40% faster computation compared with conventional models, while requiring as little as 30% of the data for adaptation. The results demonstrate a scalable, cost-efficient solution for Industry 4.0 environments with dynamic illumination and obstacles.

Abstract

Accurate indoor localization is crucial in industrial environments. Visible Light Communication (VLC) has emerged as a promising solution, offering high accuracy, energy efficiency, and minimal electromagnetic interference. However, VLC-based indoor localization faces challenges due to environmental variability, such as lighting fluctuations and obstacles. To address these challenges, we propose a Transfer Learning (TL)-based approach for VLC-based indoor localization. Using real-world data collected at a BOSCH factory, the TL framework integrates a deep neural network (DNN) to improve localization accuracy by 47\%, reduce energy consumption by 32\%, and decrease computational time by 40\% compared to the conventional models. The proposed solution is highly adaptable under varying environmental conditions and achieves similar accuracy with only 30\% of the dataset, making it a cost-efficient and scalable option for industrial applications in Industry 4.0.

Paper Structure

This paper contains 15 sections, 11 equations, 9 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System model and problem formulation
  3. Source and Target Domains
  4. Loss Function
  5. Fine-Tuning
  6. Algorithm development and implementation
  7. Pre-TL Model Development
  8. TL Algorithm Development and Implementation
  9. Performance Evaluation and Discussion
  10. VLC Measurement System
  11. Data Collection
  12. TL Model Evaluation under different lighting conditions
  13. TL model evaluation with limited data
  14. Conclusion
  15. Acknowledgment

Figures (9)

  • Figure 1: Proposed Framework of TL-based localization process.
  • Figure 2: (a) The concept of the transceiver deployment and grid distribution, and (b) the location of the transmitter and measured area in the production line.
  • Figure 3: Localization Error.
  • Figure 4: Success Rate.
  • Figure 5: Energy Consumed.
  • ...and 4 more figures