A vehicular ad-hoc network (VANET) consists of groups of moving or stationary vehicles connected by a wireless network. Intelligent vehicle networks are utilized for safety and commercial communications between vehicles or between a vehicle and a roadside unit.

VANET provides the quickest response and accurate decision-making in an emergency. Deep learning techniques are used to acquire and track the dynamics of vehicular environments, automatically make decisions regarding vehicular network traffic control, transmission scheduling and routing, and network security, and perform intelligent network resource management.

The deep learning model possesses a powerful solution to intelligent transport systems by handling a huge amount of data and knowledge extraction from a complex system. Convolutional neural networks, recurrent neural networks, autoencoders, general adversarial networks, and long short-term memory are the deep learning techniques employed in the intelligent transport system.

Some of the application tasks of intelligent transport systems using deep learning models are traffic characteristics and incidents, vehicle ID, traffic signal timing, ride-sharing and public transportation, visual recognition tasks.

Future research areas are the intelligent transport system for heterogeneous data, automatic control of vehicle action based on robust detection models with security and privacy, prediction based on various types of road traffic datasets and fusing multiple datasets, and many more.