Research And Optimization Of The Cluster Based Routing Protocol In Ad-Hoc Network

Mobile Ad Hoc network is a Mobile Multi-hop packet radio which isn't dependent on fixed infrastructures and has no central entity. Every node of this network is not only a communication terminal host, but also a router. Its main features are the equal rank of all nodes within the network, support mobility of node, networking flexibility, distributed control and easy to carry out quickly. However, because of characteristics of the Ad Hoc network, such as limited energy (power supply of nodes is generally battery), frequent changes of network topology and limited link bandwidth, the traditional network routing protocol can not be directly applied to Ad Hoc networks.Early Ad Hoc network commonly used planar structure which is characterized by simple structure, the equal status of each node, easy maintenance. But this structure limited the expansion of network. At present, the development of Ad Hoc network towards hierarchical structure. The hierarchical structure can be extended flexible. And this structure is characterized by routing and control overhead and easy network management. But how to grade the network, which is looking for the superior performance of clustering algorithm, is one of the problems of hierarchical structure currently.This paper studies the current classical clustering algorithm further. And on this basis, this paper proposes a new multi-headers clustering algorithm, Clustering Algorithm Based on Honeycomb (CABHC). And a new clustering route protocol called Cluster Based on Honeycomb Protocol (CBHCP) is advanced; which can improve performance in larger and mobility strong Ad Hoc network. The core idea of CBHCP is that the whole network divided by regular hexagon like honeycomb based on the location information of nodes, then cluster further. Select some regular hexagon like honeycomb as areas which generate cluster headers, and select some hexagon which is are next to those areas as their members. The strategy that is multi-header in this algorithm solves the bottleneck problem brought by previous single cluster header (such as CBRP). It eases heavy burden on a single cluster header, achieves the burden sharing and enhances the stability of network. At last, make CBHCP and CBRP simulation experiments in multiple test environments on NS2, compare the results analyzed and prove CBHCP is feasible and superior.