Nowadays, with the expansion of the network and its applications, including wireless sensor networks and the Internet of Things, various types of smart systems and services are introduced to increase the quality of human life. For instance, we can point out the telecommunication between cars in intelligent transportation systems in order to form a network for collecting traffic information and providing services such as routing, traffic control and etc. One of the methods for establishing communication between cars, in which cars transfer information independently and without the need for a separate network infrastructure, is the vehicular ad-hoc network (VANET) which is a subcategory of mobile ad-hoc networks. Because of using no dedicated network infrastructure, vehicles should be able to communicate with other vehicles independently. Also, each car is charge of routing message packets in this network by itself. Of course, the main feature of VANETs, which distinguishes them from other mobile ad-hoc networks, is higher movement speeds, sufficient access to the source of energy and predictable paths of the nodes. Several protocols have been presented so far to overcome the challenges in these networks such as fair media channel allocation, real-time routing, security and etc. But, due to high speed of vehicles and also their continuous speed change inside a highway, VANETs are facing the challenge of instability of network topology which makes it difficult to maintain good communication in this network. Therefore, the stabilization of VANETs is the most important challenge in these networks, and solving it can help to solve other challenges as well. One of the methods for stabilizing these networks is clustering and choosing the appropriate cluster head for each cluster in a way that helps to increase the stability. Choosing the appropriate cluster head is a challenging problem which we focus on it in this research. We try to investigate the inherent characteristics of VANETs and parameters affecting the stability of these networks. Also, previously presented protocols for choosing the appropriate cluster head in VANETs are examined to identify and eva‎luate their strengths and weaknesses different conditions. As the main contribution, according to the car movement models and the laws governing the highways, a new clustering method is presented in which, more comprehensive parameters that describe the spatial changes of the cars in these networks are taken into account during the clustering phase. So more appropriate nodes will be selected as cluster heads to increase the stability of the networks. This is carried out regarding the speed difference, average distance and duration of connection among the neighbors of each node. Then using a fuzzy logic system, a specific weight will be assigned to each car based on these parameters to represent its suitability for being clusterhead. Simulation results of utilizing the proposed method in different speed ranges on the highway, show a decrease of approximately 31.3% in the production rate of new cluster heads and 20.8% in the average number of status changes per cars compared to the previous works. This indicates more stability of the clusters created by the proposed method.

امير خورسندي

عليرضا بصيري، زينب زالي