Optical camera communications for IoT applications

• Main Author: Fernandes, Dario Duarte
• Publication Date: 2023
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10773/41695
• topic: IoT; OCC; YOLO; 2D visual codes; Camera; Object detection

The Internet of Things (IoT) has become increasingly present in our daily lives, revolutionizing the way we interact with technology. With the emergence of 5G, the number of connected devices has skyrocketed, further amplifying the importance of IoT. However, this rapid growth has raised concerns about the limitations of radio frequency (RF) technologies and their ability to handle the increasing number of devices within the limited available spectrum. As more devices connect to the IoT, the demand for other solutions with potential to alleviate the RF constrains becomes crucial to ensure seamless connectivity and prevent signal interference. This dissertation describes the development of a communication system using Optical Camera Communication (OCC), which transmits information through visual codes captured by a low-cost camera module connected to an embedded system. The visual codes require initial image processing for reception and decoding of information. This enables the exploration of advanced Deep Convolutional Neural Networks (D-CNNs), like YOLOx, through a comprehensive process of training, validation and testing with custom-designed 2D visual codes. Two types of symbols are introduced. The first type prioritizes accurate location, including features like an orientation indicator and symbol identifier, reducing the available space for data. The second type prioritizes data rate, utilizing the entire symbol exclusively for conveying data. The research findings underscore OCC’s potential as a viable solution to address RF limitations within the IoT landscape. Despite the system’s lower speed performance, it demonstrated effectiveness with a low error rate, indicating a successful implementation and evaluation of OCC using D-CNNs and custom visual codes. This showcases the system’s capacity to provide efficient, reliable, and cost-effective communication for IoT applications. The results open the door for further improvements and in-depth studies, positioning OCC as a promising alternative to traditional RF technologies.