Developing An Automatic Road Safety Model for Accident Identification, Detection, And Prevention Using Deep Learning Algorithms

Abstract

Road traffic monitoring systems are one of the leading real-time applications that use the Internet of Things to monitor and identify traffic scenarios on the road. IoT devices are integrated with sensors to sense and capture data within a specific distance range and transmit it to other sensors within a coverage range. Communication is possible, and data will be passed from one device to another only if they are located within the sensing and coverage region. Thus, multiple IoT devices are interconnected logically and communicate with one another within a coverage region. This paper focused on creating an efficient IoT network to monitor and broadcast accidental information immediately to the other vehicles on the road at a defined distance. Some exciting works included installing CCTV cameras, IoT devices, and other sensors only on road junctions and signals, where they can monitor only at particular locations, and they are not efficient in accident detection over urban city roads. This paper has focused on deploying more IoT devices within an urban city and creating an IoT network for accident detection and prevention. The IoT data are analyzed using a robust and efficient deep learning model, Convolution Neural Network work, that can quickly predict accidents from the IoT data analytics and intimate to the admin to broadcast the message to all the vehicles and the users on the road to take prevention actions. The IoT data is analyzed using the CNN algorithm implemented in Python, and the results are verified. The performance of the proposed CNN model is evaluated by comparing its output with the other state-of-the-art methods and proving that CNN outperforms the others.