Poly-parametric performance evaluation of mobility models for clustering Wireless Mobile Sensor Networks

摘要

In wireless mobile sensor networks (WMSNs), the majority of clustering algorithms employ random waypoint (RWP) to consider mobility in nodes. In this work we evaluate the performance of clustering under RWP and random walk (RW) derivatives mobility models through simulations, since a particular mobility model cannot represent the mobility behavior of all nodes. The performance evaluation is conducted for two clustering approaches; the non-overlap and overlap clusters. In the first, a non-cluster-head node belongs exactly to one cluster, while in the second to more than one cluster. The latter is quite realistic and scarcely examined in the literature. Simulation results show that the performance of a clustering algorithm is strongly affected by the node mobility model selection and by many aspects of the model.