Artificial intelligence routing algorithms in inter-vehicle mobile networks

Author: 
Mohammad Ordouei, Amin Shams and Mastooreh Moeini

Considering the development of communication technology and the dependence of today's society on it, its use in transportation is important. The increase in many problems in urban traffic management, driving and guidance of cars on suburban roads and the heavy financial and human costs caused by these problems have become a big challenge. At the same time, the development of communication and information technology as an advantage clearly shows the need to use intelligent vehicle communication technology in the field of transportation. VANET is one of these technologies proposed in the vehicle transportation network. One of the problems of inter-vehicle networks is routing or packet transfer speed, which routing is done using mobile network. This network is a protocol based on the TORA protocol that performs routing whenever needed. In fact, it is a demand-based protocol, which means that the source node performs a routing before sending the packet, and a route is not created until the route is needed. This protocol is proposed for very dynamic mobile networks, which we have integrated with genetic algorithm to improve packet sending delay. After the initial routing by the TORA protocol, a population of existing routes is created, which we call the initial population in the genetic algorithm. By using the function of fitness, the main path for the transmission of packets is determined from among the created paths. The position of the cars is determined, and then the relationship between the positions is determined, and in the following section, it is determined that the genetic algorithm has shown its efficiency and increases the speed of data transmission in routing. In this context, the radio position is important, it is determined by the necessary forecasts. There is an error in the setting of the radio system because there is also a disconnection. In general, this system can be implemented with a small error percentage.

Paper No: 
4643