Reinforcement Learning-Based Routing Protocols in Vehicular Ad Hoc Networks for Intelligent Transport System (ITS): A Survey

Today, the use of safety solutions in Intelligent Transportation Systems (ITS) is a serious challenge because of novel progress in wireless technologies and the high number of road accidents. Vehicular ad hoc network (VANET) is a momentous element in this system because they can improve safety and efficiency in ITS. In this network, vehicles act as moving nodes and work with other nodes within their communication range. Due to high-dynamic vehicles and their different speeds in this network, links between vehicles are valid for a short time interval. Therefore, routing is a challenging work in these networks. Recently, reinforcement learning (RL) plays a significant role in developing routing algorithms for VANET. In this paper, we review reinforcement learning and its characteristics and study how to use this technique for creating routing protocols in VANETs. We propose a categorization of RL-based routing schemes in these networks. This paper helps researchers to understand how to design RL-based routing algorithms in VANET and improve the existing methods by understanding the challenges and opportunities in this area.