Design and Implementation of Deep Learning-Based MIMO C-OOK Scheme for Optical Camera Communication Considering Mobility Environment

Currently, wireless communication systems using radio frequency are widely utilized across various applications for example: mobile communication, satellite networks, and Industrial Internet of Things (IIoT). Wireless communication, owing to its convenient installation, offers advantages over wired counterparts. Nevertheless, utilizing high frequencies for wireless data transfer raises substantial concerns for human health. To address this issue, researchers have explored alternatives deploying visible light in communication systems, leading to three relevant approaches: visible light communication, light fidelity, and optical camera communication. The role of artificial intelligence (AI) in industries and society is pivotal, as it addresses intricate problems, devises intelligent resolutions, and gradually replaces human intelligence in driving emerging technologies like smart home, data analysis, and IIoT. In this study, we introduce MIMO C-OOK architecture, leveraging a convolutional neural network for light sources tracking and a deep learning algorithm for decoding. The proposed method is designed to enhance the performance of the traditional C-OOK such as improving data rate, higher distance, and reducing BER. This approach aims to attain a 8 meters distance with good bit error rate in 3 m/s mobility environment. This is achieved through meticulous control of shuttle speed of camera, camera focal length, and the implementation of suitable FEC.