A methodology for designing of intelligent services in ubiquitous communications

Mobile ad hoc networks (MANETs) are recognised as one of the key infrastructures in ambient intelligence (AmI) due to their promising ubiquitous connectivity. Key factors for reliable performance of service protocols in MANETs are the manner in which they adapt to route changes caused by mobility and communicating traffic patterns. Our main idea to support AmI ubiquitous communications is to use fuzzy C-mean clustering as data mining technique to extract knowledge and dependencies from high-level information of the application scenario. We use the resulting knowledge to predict the upcoming context and make MANETs services context-aware. Thus, a mobile node builds dynamic models determined by IF-THEN rules for its various environments and adapts its service protocols. We propose a methodology for storing and processing MANETs context, online fuzzy reasoning about it and improving the performance of underlying service protocols based on simple distributed, unobtrusive and handling different scenarios intelligent algorithm.