Convolutional neural networkscheme–based optical camera communication system for intelligent Internet of vehicles. (April 2018)

• Record Type: Journal Article
• Authors: Islam, Amirul; Hossan, Md Tanvir; Jang, Yeong Min
• Is Part Of: International journal of distributed sensor networks. Volume 14:Number 4(2018)
• Publication Date: 2018-04
• Subjects: Optical camera communication -- Internet of vehicles -- convolutional neural networks -- vehicular communications; Sensor networks -- Periodicals; Intelligent agents (Computer software) -- Periodicals; Multisensor data fusion -- Periodicals; 681.2
• DOI: 10.1177/1550147718770153
• Languages: English
• ISSNs: 1550-1329

Abstract

The evolution of the Internet of vehicles and growing use of mobile devices has created a demand for new wireless communication technologies. Optical camera communication, which uses light-emitting diodes as transmitters and cameras as receivers, has emerged as a promising alternative. Since light-emitting diodes and cameras are already exploring in traffic lights, vehicles, and public lightings, optical camera communication has the potential to intelligently handle transport systems. Although other technologies have been proposed or developed in both academia and industry, they are not yet mature enough to uphold the huge requirements of the Internet of vehicles. This study introduces a new intelligent Internet of vehicles system based on optical camera communication combined with convolutional neural networks. Optical camera communication is a promising candidate for maintaining interference-free and more robust communication, for supporting the Internet of vehicles. Convolutional neural network is introduced for precise detection and recognition of light-emitting diode patterns at long distances and in bad weather conditions. We propose an algorithm to detect the interested light-emitting diode signals (i.e. regions-of-interest), measure the distance using a stereo-vision technique to find out the desired targets, and simulate our proposed scheme using a MATLAB Toolbox. Thus, our system will provide great advantages for next-generation transportation systems.