An IoT aware Road Accident Prevention System for Smart Cities Using Machine Learning Techniques

An IoT aware Road Accident Prevention System for Smart Cities Using Machine Learning Techniques

S. Deeksha, Sannasi Ganapathy, V. Muthumanikandan
Copyright: © 2023 |Pages: 15
DOI: 10.4018/978-1-6684-7756-4.ch001
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

In smart cities, road accidents are very serious and avoidable. In addition, road accidents are very common in India, and they are also becoming a serious issue all over the world. In this world, the people are not able to reach hospitals due to the traffic, lack of transport facility, and unavailability of hospitals after the accidents. To enrich the population of urban people, the “smart cities” are developed to enhance the sophistication in daily life through technological development. The proposed image classifier model has been tuned with different values of hyperparameters such as number of units, activation function, optimizer, learning rate, and number of epochs. The efficiency and accuracy of the model is duly considered while building the model for predictive analysis. The images were transformed and augmented before feeding them into the neural network to ensure proper training by blocking over-fitting of the model because of lack of data. The proposed model achieves 98.21% accuracy, which is greater than the existing works.
Chapter Preview
Top

Literature Survey

Many accident prediction systems are available for predicting the accidental area in the various regions of the world. Among them, Kinoshita et al. (2022) paper deals with a dataset which shows certain regions in Japan with heavy snow conditions which lead to traffic congestion. A laser-based method and depth map based on snow piled in the corners. The benefits of this solution include that piled snow on road shoulders is taken into account, with the capacity of detecting 3 levels of snow conditions. However, it is only useful in snowy conditions and is unidirectional. Abramowski et al. (2018) paper identifies the cause of accidents based upon DVRs and surveillance cameras. This helps identify formulas devised for speed, acceleration and energy of vehicles. The solution proposed is kind of a black box solution to know the cause of accidents and not completely avert them, guidelines can be enforced based on outcomes of research. Rajendran et al. (2022) use ESP and USB camera together with the help of cNN detect potholes on highways and roads and send the information to the highway department this help detect potholes with a notification service for the ministry of road transport and highways to increase the efficiency of their functioning.

Complete Chapter List

Search this Book:
Reset