Vehicular network structures present a range of challenges and opportunities for efficiently managing awareness of road dynamics and network connectivity. An enhanced manageable organization can offer a better reaction to safety-related road events, facilitate dynamic topological flexibility, relate to road layout, and interact with unpredictable distribution of the vehicles. Vehicular grouping is one of the suggested structural techniques that offers a great benefit in grouping vehicles and modelling data routing, giving importance to road structure and the occurrence of a dynamic event within the associated group of vehicles. The approach discussed in this paper is based on a dynamic grouping through phases of self-formation, self-joining, self-leaving and self-healing as key components of the protocol operational cycle. Both vehicular physical connected resources and the remote computational cloud could be used for data processing and monitoring of road dynamics. This, in effect, encourages an Internet of Things (IoT) environment that enhances the dynamic performance through direct interaction between the virtualized network of vehicles and the physical network on the road leading to Internet of Vehicles (IoV). The objective of this paper is to develop a concept of network self-formation algorithm based on vehicle grouping strategy wherein the node can flexibly switch its function, be it an IoT gateway or a router node, based on the proposed fitness election model to be elected as group head. Testing using Contiki-Cooja simulator has been implemented on various road condition scenarios reflects the operational ability of the algorithm taking into consideration the network performance based on the ultimate capacity of the road.