Intelligent Approaches for Routing Protocols In Cognitive Ad-Hoc Networks

Abstract

This dissertation describes the CogNet architecture and five cognitive routing protocols designed to function within this architecture. In this document, I first provide detailed modeling and analysis of CogNet architecture and then provide the detailed approach, mathematical analysis, and simulation results for each of the developed cognitive routing protocols. The fundamental idea for these cognitive routing protocols is that a proper and adaptive network topology should be constructed from network nodes based on predictions using cognitive functions and past experience. The nodes in the cognitive radio network employ machine learning techniques to use past experience and make wise decisions by predicting future network conditions. The cognitive protocol architecture is a cross-layer optimized construct where the lower layer knowledge of the wireless medium is shared with the network layer. This dissertation investigates several intelligent approaches for cognitive routing protocols, such as the multi-channel optimized approach, the scalability optimized cognitive approach, the multi-path optimized approach, and the mobility optimized approach. Analytical and simulation results demonstrate that network performance can be increased significantly by applying cognitive routing protocols.