Fuzzy Rule-Based Mobility and Load Management for Self- Organizing Wireless Networks

Abstract

Mobility management in a cluster-based, multihop ad hoc network is studied. It is shown that the process of clustering the network into groups of stations has similarities to data analysis, in particular, pattern recognition. In data analysis, the term clustering refers to the process of unsupervised learning, which also describes the situation in a mobile ad hoc network. In this paper, existing data-clustering algorithms are first classified into different categories. Some of the most important types of algorithms are afterwards described, and their applicability to the problem of mobility management in an ad hoc network is studied. It is shown that most of the pattern-recognition algorithms are not suited to the application under consideration. This is why we have developed a new clustering scheme that incorporates some of the ideas of the data classification schemes. The new clustering scheme is based on a rule-based fuzzy inference engine. The main idea consists of the consideration of dynamic clustering events chosen as a consequence of the fuzzy rules. Four types of clustering events are considered. The performance of the clustering algorithm has been evaluated by computer simulation.