Adaptive Modular Fuzzy-based Handover Decision System for Heterogeneous Wireless Networks

Future generation wireless networks will demand more intelligent and adaptive handover decision mechanisms to fulfil users’ expectations in terms of seamless mobility over extensive area, minimum price, high data rate, adequate QoS provision and so on. For such a demanding networking environment the handover decision system (HDS) need to be highly intelligent. Fuzzy logic appears to be one of the methods that can be employed to enhance HDS intelligence. However, most existing fuzzy-based HDS designs proposed in the literature are monolithic, i.e. based on a single fuzzy engine. In view of the growing demand for real-time applications, it is becoming necessary to include a relatively large number of decision parameters (especially QoS-related) in the handover decision process. However, an increasing number of decision parameters give rise to generating a very large number of fuzzy rules. This in turn increases computational complexity and requires significantly long algorithm execution time, which may not be acceptable for real-time applications. Furthermore, if the same fuzzy membership functions (FMFs) and fuzzy rules are used for all traffic types (e.g. VoIP, video streaming etc), the HDS may not give the overall best decision results, as each traffic type has a different set of QoS requirements. In order to address the above issues, we are proposing a modular design concept to deal with the algorithm execution time and an adaptive mechanism to select FMFs and fuzzy rules which are best suited to the incoming traffic. The results show that the modular design significantly reduces the algorithm execution time while an adaptive mechanism improves the network selection performance considerably.