Loading...

Machine Learning-Based Positioning in Optical Communication Networks Using a Camera: Indoors and Underwater

Seyed Tabatabaei, Raouf | 2021

405 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54129 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Shabany, Mahdi; Hashemi, Matin
  7. Abstract:
  8. Positioning refers to the process of estimating the receiver coordinates aiming for an accurate understanding of the surrounding environment. This branch of science has attracted considerable attention in recent years. Today, the influence sphere of positioning is so expanded that encompasses applications in daily life (e.g., navigation) as well as commercial and even military fields. Global Positioning System (GPS) is the most widely used tool in real-time positioning. Because GPS imposes great measurement errors in challenging conditions (e.g., turbulent water environments), alternative methods such as methods based on Wi-Fi, Bluetooth, and visible light have been proposed. Benefiting from LED lamps and their advantages including broadband, high security, low power consumption, long life-time, cost-effectiveness, and low maintenance, the visible light telecommunication network is considered one of the pioneers in this field. Despite the visible light’s potential for indoor positioning, existing methods have not yet reached optimal maturity owing to the excessive errors. The lack of existing research in this field has led us to fill the gap by achieving an algorithm for accurately calculating the position and orientation of the receiver in both indoor and underwater environments. Although this is possible using a variety of methods, a camera-based geometric approach with higher accuracy, lower cost and capability of commercialization is the best matching. This study aims to provide an accurate model for simulating underwater images and a geometry-based algorithm to calculate the position and orientation of the receiver (e.g., camera) in an optical communication network. Due to practical limitations to use the camera in harsh environments (especially in water), the design, testing, and evaluation of the process were entirely simulated. Through the present study, all positioning algorithms outside the water environment were carefully investigated. A novel algorithm based on geometric calculations was then obtained to find the position and orientation in an optical network. Next, by examining the water channel's characteristics, a model of making realistic underwater images was presented. Since some distortions might occur in the simulated images, an appropriate neural network was created and trained by machine learning algorithms to extract basic information from obtained images. Finally, the proposed model was simulated in a synthetic aqueous environment to evaluate its accuracy and performance. It is worth noting that all codes and simulations were written by the Python programming language
  9. Keywords:
  10. Optical Communication Networks ; Simulation ; Machine Learning ; Neural Network ; Localization ; Underwater Environment ; Enclosures

 Digital Object List

 Bookmark

No TOC