Vehicular Ad-Hoc NETworks (VANETs) have received considerable attention in recent years, due to its unique characteristics, which are different from Mobile Ad-Hoc NETworks (MANETs), such as rapid topology change, frequent link failure, and high vehicle mobility. The high mobility in VANETs causes high topology changes and in turn leads to excessive control overhead messages and frequent link failures.
Traditionally, clustering techniques have been used as the main solution to reduce the control overhead messages in VANET, in which the network is divided into multiple clusters and selecting one of the Cluster Members (CMs) as a Cluster Head (CH). The selected CHs are responsible for coordinating the members of the cluster, and communication between clusters. The clustering techniques will significantly reduce the routing control overhead messages, that is because the clustering techniques restrict the communication between each CM and it’s CH instead of communication between all the vehicles in the VANET topology.
The most important characteristic for any clustering technique is to create a stable cluster with minimum clustered control overhead messages. In this thesis, we define four types of clustered control overhead messages, as follows: The control overhead messages generated due to clustering formation and maintenance, the control overhead messages due to forwarded from the CMs to the CH, the periodically broadcasted advertisement messages by the CH, and the control overhead messages broadcasted due the CH election process. Also, we assume all VANET networks are already preclustered. Therefore, the clustered control overhead messages generated by the clustering formation or maintenance are eliminated. Furthermore, the clustered control overhead messages generated by the CM and the CH, and due to CH election processes still produce high clustered control overhead messages.
In the first part, the problem of Clustered-Based Routing (CBR) protocol in VANET is relaying the received data to unstable CH. Most of the proposed CH election algorithms select the CH in the clustered VANET topology by considering many parameters, such as: mobility, location, and node Life-Time (LT). Most of the clustering techniques produce unstable clustered VANET topology, because they do not elect the stable CH. Therefore, the cluster that elects the suitable CH which does not change frequently is considered a stable cluster. Also, it improves the CBR protocol performance in terms of throughput and end-to-end delay. Therefore, we propose a Cluster-Based Life-Time Routing (CBLTR) protocol. The CBLTR protocol aims to increase the route stability and average throughput in a bidirectional segment scenario. The Cluster Heads (CHs) are selected based on maximum Life-Time (LT) among all vehicles that are located within each cluster. We propose also an Intersection Dynamic VANET Routing (IDVR)protocol. The IDVR protocol aims to increase the average throughput, and to reduce end-toend delay in a grid topology. The elected Intersection CH (ICH) receives a Set of Candidate Shortest Routes (SCSR) close to the desired destination from the Software Defined Network (SDN). The IDVR protocol selects the optimal route based on its current location, destination location, and the maximum of the minimum average throughput of SCSR.
In the second part, a problem occurs when the control overhead messages increase due to periodically forwarding of CM HELLO (CMHELLO) messages between the CMs and the CH, and when the CH periodically broadcasts an CH ADvertiSement (CHADS) messages to declare itself to the CMs. Hence, minimizing control overhead messages in any cluster environment is an essential goal to use the network resources efficiently. Therefore, we propose two algorithms: First, a Control Overhead Reduction Algorithm (CORA) which aims to reduce the control overhead messages in a clustered topology, by developing a new mechanism for calculating the optimal number of CMHELLO messages. Second, an Enhanced version of CORA (ECORA) which aims to reduce the CHADS messages that broadcasted by the CHs, by proposing a CHADS prediction algorithm that enables the CH to predict the period of time for broadcasting the CHADS messages.
In the third part, the frequent CH election mainly increases the clustered control overhead messages, which yields to consume high amount of available network resources. High clustered control overhead messages is considered as main problem that negatively impacts the network performance. In this part, we concentrate on the reduction of CH election control overhead messages. Therefore, we propose a new Passive CH election avoidance (PCHEA) protocol that aims to optimize the number of CH election process. In PCHEA protocol, each CH selects another CH based on specific information already stored in its memory, without requiring to trigger the election function. The CH sends to the CMs the next CH identification and its activation time. Also, we propose a CH Routing (CHR) protocol that aims to reduce the number of relayed CHs between any pair of vehicles. In CHR protocol, the CH selects the second adjacent CH among all CH located within its transmission range. The PCHEA protocol and CHR protocol significantly reduce the number of CH elected and increase the average throughput in a bidirectional highway scenario, respectively.
Finally, the proposed protocols are evaluated using SUMO version 0.28.0 traffic generator and MATLAB version R2016b. We compare the performance of the proposed protocols with other protocols in the literature, in different scenarios and in terms of different performance metrics.
| Date | 10 Nov 2018 |
|---|
| Original language | American English |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Michel Kadoch (Supervisor) |
|---|