Date of Award

7-2006

Degree Name

Doctor of Philosophy

Department

Computer Science

Abstract

In recent years, wireless networks have enjoyed tremendous development and popularity due to the technological advances of wireless radio devices. While interest in conventional wireless networks continues to evolve, new types of wireless networks, such as Mobile Ad Hoc Networks (MANET) and WirelessSensor Networks (WSN) are evolving fast and receiving much attention from academia, industry, and government.

Scalability and energy-efficiency present two of the most important challenges in Mobile Ad Hoc Networks. Scalability in MANETs can be defined as the capability of the network to provide an acceptable throughput when the network size increases. Energy-efficiency can be defined as using less energy to deliver greater or equal amount of services. MANET nodes rely on portable, limited power resources and they bear the duty of relaying packets for other nodes. The failure of some nodes due to battery drainage might bring the whole network down.

In this dissertation, we propose clustering and routing approaches to overcome scalability and energy constraint problems that exist in large scale MANETs. To handle these problems, we design a hierarchical energy-efficient scheme that can be easily setup and maintained. The scheme has three major components, namely: (1) the hierarchical (clustering) component, (2) the maintenance component, and (3) the routing component. Our research provides a mathematical formulation of the clustering component using integer linear programming---ILP. It also provides centralized and distributed energy-efficient schemes that achieve efficient clustering, efficient maintenance, and efficient routing in large scale MANETs. Our simulation study shows that our suggested approaches outperform well known schemes found in the literature.

The distributed energy-efficient scheme presents one of our main contributions. The distributed scheme builds and maintains a hierarchical network structure using low message and time complexity. Furthermore, it provides an intelligent path selection controller that can be easily incorporated in any existing link state routing protocol to select energy-efficient routes. The hierarchical structure increases scalability, but does not improve energy-efficiency. To achieve energy-efficiency, we propose three fuzzy logic controllers that aggregate multiple network metrics and then decide on the quality of the node/path. This quality is then used as an important parameter in the clustering, maintenance, and routing components. The first controller is used to achieve efficient hierarchical design, the second controller is used to achieve efficient network maintenance, and the third controller is used to achieve efficient routing.

Access Setting

Dissertation-Open Access

Share

COinS