Dirk Pesch, Susan Rea

Fuzzy Logic based Multi-Metric Route Management for Ad-hoc Networks using Realistic Mobility Mode

Cork Institute of Technology (Ireland)

Routing is an essential function for traffic management in ad hoc networks. Table-based routing algorithms used in proactive protocols are computationally intensive and require periodic transmission of status information amongst all network nodes. Routing table sizes can also swell with network size and can be large for dense networks. In order cope with possibly rapidly changing topology reactive protocols such as Dynamic Source Routing (DSR) generate routes for unknown destination paths on an as needs be basis. So as to avoid the continual demand for route discovery in DSR, network nodes record previously determined routes in either path or link caches. A route discovery attempt can possibly result in several paths being uncovered for a single destination. As nodes often have a finite capacity path cache, it may not be possible to store all paths. In order to influence productive caching decisions a fuzzy logic system is applied to the route discovery technique to curb nonoptimal network floods. This action causes a cessation in the generation of low quality routes as only paths with good routing metrics are selected for the rebroadcast of route discovery packets. Consequently, route query packets arriving at the necessary destination node, or at some intermediate node with knowledge of the destination node, generate high quality route replies. Furthermore, node mobility compounds the problem of route maintenance and accurate modeling of node mobility is essential in designing effective route cache management strategies. This paper presents a fuzzy logic based decision algorithm that weighs individual links as a path to the necessary destination is being constructed if this link is deemed suitable by the fuzzy logic system it is added to the path and route construction continue. The fuzzy controller is used to instruct caching decisions and to optimize route selection as only good quality links are recorded in source destination paths. Three mobility models are applied to the analysis of the proposed fuzzy controller in order to demonstrate the importance of proper mobility modeling on cache management techniques.