Location Information Based Clustering Algorithm

With the growing demand for communication and the popularity of variousportable devices, mobile ad hoc networks has been widely used on the battlefieldcommunications, Internet of Things, disaster and emergency relief, mobile office andso on. As the number of nodes increases, clustering structure is used to enhance thescalability of mobile ad hoc networks. Following study shows that the clusteringstructure can not only enhance the network scalability, but also is a great help fornetwork resource management such as channel access, routing, network security andenergy control. It has great significance for network performance.Currently, the speed of the nodes increase a lot, and network topology changesmore rapidly. This brings new challenges to the clustering algorithm. Consequently, anumber of clustering algorithms for high-speed mobile environment are invented. Byputting the nodes with similar speed into one cluster, the new invented algorithmsprolong the survival time of the node and enhance the stability of the clusteringstructure in high-speed mobile environment. However, these algorithms are generallymore complex. And overhead for clustering and maintenance is very huge. Eachcluster overlaps too much space, making it difficult for frequency planning, which ismore serious in mobile environment.Taking both clustering and frequency planning into consideration, a locationinformation based clustering (LIBC) algorithm is proposed. Location information isused to calculate the observation speed. Kalman filter is used to make it moreaccurate and enhance the robustness of the network. By introducing virtual networkcentral node to reflect the group movement of the network and using the relativepositions to form the cluster area, LIBC extends in-cluster survival time of the nodes.Also, by early warning of cluster head failures and replacing cluster heads in advance,LIBC is able to reduce network jitters. And by frequency planning and load balancingbetween clusters, LIBC reduces the collisions between the nodes.To verify the performance of LIBC, a corresponding communication simulationsystem is designed, and a series of simulation experiments are conducted. Simulation results show that the location information based clustering algorithm not onlysucceeded in forming a stable and reasonable clustering structure but also was easyfor frequency planning, significantly improves network performance in terms ofthroughput, packet loss rate and delay, and is suitable for a multi-node, middle or highspeed mobile environment.