Fuzzy-control-based clustering strategy in MANET

Describing briefly the characteristics of a mobile ad-hoc network, comparing with related works and considering the highly dynamic topology and absence of any fixed infrastructure in a mobile ad-hoc network, a new time degree definition was given. By analyzing the link relationship of the neighboring nodes, a network model based on the link life of two nodes was created. The use of fuzzy control techniques was investigated. For each metric, a fuzzy membership function was defined to predict a more stable link. A fuzzy-inference rule base was implemented to generate the fuzzy cost of each link. A d-degree clustering algorithm based on a mobility prediction scheme was proposed in a distributed manner. Simulation results show that the proposed scheme outperforms LCC and HD clustering scheme.