International Journal of Engineering Applied Sciences and Technology, 2022 Vol. 7, Issue 6, ISSN No. 2455-2143, Pages 316-324 Published Online October 2022 in IJEAST (http://www.ijeast.com) 316 LI-FI BASED VEHICULAR COMMUNICATION NETWORK WITH ACCIDENT PREVENTION AND DRIVER ALERTING SYSTEM K. A. R. D Perera, T. R Amarasena, L. H Edirisinghe, Dr. Anuradha Jayakody, N. K. A Dissanayake, Ms. Wellalage Sasini Nuwanthika Department of Computer Systems Engineering Sri Lanka Institute of Information Technology, Malabe, Sri Lanka Abstract— Many vehicle accidents were recorded because of the drowsiness or emotional instability of the drivers. There are many fatal cases and minor cases are being recorded every day and these numbers are on a gradual increase. Some vehicle manufacturers design several safety measures such as air bags, powerful roll cages, automated braking systems to mitigate the damage of the accidents. But always prevention is better than cure. There are occasions which accidents occur due to the malfunction of the vehicle, but most of the accidents are because of the human errors done by the driver. And these human errors can be happened by the driver inside the vehicle or any driver who is driving on the road. There can be a chance of preventing such accidents if the danger could be communicated earlier to the driver. Therefore, not only detecting the current state of the driver but also communicate such a situation to nearby vehicles is very important. The proposed solution of this report is a hardware-based system which explore the ability to detect the drowsiness of the driver using image processing technologies to prevent vehicle accidents by alarming the driver and as well as communicating the danger to the nearby vehicle using light fidelity technology. Keywords— Li-Fi, formatting, Accident prevention, Drowsiness Detection I. INTRODUCTION Road accidents are much prominent now days due to higher numbers of vehicles on the motorways. These accidents have caused minor and as well as fatal injuries to the drivers and passengers. There are approximately 1.3 million deaths reported due to road accidents on the year 2021. According to the previous records the number of deaths is on the rise. Therefore, researchers are constantly involved in innovating solutions to minimize these accidents. Many technologies such as drive assistance, lane departure warning and proximity warning have been developed and implemented to the vehicles. These technologies used many detection media such as Wi-Fi, Bluetooth, infrared, and ultra-sonic waves. These technologies have their own strengths and weaknesses. Proposed light fidelity-based vehicle to vehicle and vehicle to infrastructure technology warns the drivers of the vehicles to avoid a potential vehicle accident thus minimizing the accident completely or reducing fatality of the accident. Light Fidelity (Li-Fi) is a recently discovered low cost and less complex technology which has a higher data transmission speed. Li-Fi uses visible light as the medium to transfer data. The binary data transmission from one node to the other is transmitted by using the light intensity changes. The infrastructure for the Li-Fi data transmission is already available in vehicles. The LED headlight of the vehicle will be acting as the transmitter while a separate connected module in the nearby vehicle on the line of sight will be acting as the receiver. The proposed solution consists with sub systems to identify the drowsiness of the driver, drunkenness of the driver, identification of the proximity of nearby vehicles and vehicle to infrastructure communication on controlling traffic lights to allow emergency vehicles to pass by any junction using Li-Fi technology. The solution has its own visual and audio warning system to warn the driver inside the vehicle of any of above hazards were detected. II. RELATED WORK Several approaches have been proposed related to this issue in many papers. Some specific papers have been analyzed in the following paragraphs. The face is a significant organ because of how much information it communicates. Facial expressions, such as blink and yawn rate, change when a driver is fatigued. A research paper published by several students of Beijing University of Technology propose a technology called DriCare that analyzes video footage to determine a person's tiredness level based on telltale signs like yawning, blinking, and the length of time their eyes are closed. After realizing the limitations of existing methods, we develop a brand new face-tracking algorithm to address this issue. As an additional step, we developed a novel approach of facial area recognition using a set of 68 landmarks. Then, we analyze