Adaptive power management algorithm in mobile ad hoc network

Title

Identifier

d_2009_2013:6285fadd3355:11257

identifier

11579

Creator

Pradhan, Nuraj Lal,

Contributor

Tarek Saadawi

Date

2012

Language

English

Publisher

City University of New York.

Subject

Electrical engineering | Mobile Ad Hoc Network | Power management algorithm | Topology Control | Wireless Network

Abstract

This thesis addresses the importance and the necessity of distributed power management algorithms that adapts to the changing physical environment and network topology in Mobile Ad Hoc Network (MANET). Ad hoc networks are a self-organizing systems formed by the cooperating nodes within a communication range. Routing protocols are responsible for enabling nodes in ad hoc network to communicate with each other over an error prone wireless channel. A key component that achieves and maintains connectivity between the nodes in the network is the power control algorithm. However, without a central node to administer the power control, building network topology is more challenging in wireless ad hoc networks. Further, if the ad hoc network is large consisting of thousands of nodes, then collecting information from all the nodes and passing it to the concerned nodes lead to high overheads. Thus, distributed power control algorithms that are asynchronous, scalable and localized are particularly attractive in MANET.;We propose a distributed power management algorithm (DISPOW) that adaptively manages the transmit power of nodes in MANET to preserve network connectivity. It builds a unique stable topology in a completely distributed manner tailored to its surrounding non-uniform node density and propagation environment. However, a node in order to be strongly connected can increase its power indiscriminately causing interference. Also, asymmetric links developed by distributed algorithm are a major source and interference and most of the routing algorithms do not use them to route packets. DISPOW with a receiver-based interference model not only co-operatively reduces inter-node interference but also convert asymmetric links to symmetric links if required. The adaptive optimization algorithm is shown to provide robust strongly-connected network with improved performance.;Another proposed algorithm, distributed power management algorithm with directional antenna (DDISPOW), highlights the benefits of using directional antenna with power control algorithms. Smart antenna techniques can increase channel utilization and be an effective technique in reducing interference in a network.;To demonstrate the performance and capability of the proposed algorithms, DISPOW and DDISPOW, we have evaluated them in a realistic ad hoc network through theoretical and simulation analyses. Modeling the complex propagation and fading losses a signal experiences over a dynamic wireless channel is an integral part of a realistic ad hoc network. It affects the connectivity and stability of a network and crucial parameters of different protocols. Traditional disk propagation model, free space propagation model or the 2 ray model do not model the realistic physical environment where there are several multi-paths of similar strength and the propagation range is limited by batteries. Additionally, interference from other nodes in multi-user system acts as noise and can be more significant than background thermal noise. A multi-user wireless networks are generally interference-limited rather than noise-limited.;Therefore, the log-distance path loss model, random log-normal shadowing model and the Rayleigh fading model is used to model the wireless channel in different environment such as a city or an open area. Analyzing the performance of the network over a city and a rural environment led to two distinctly different results. Higher attenuation and degradation of wireless links in urban environment induced longer packet delay and more overheads and consequently lower throughput. Further, node mobility introduces additional fading and frequent changes in topology. This change in topology has to be reported throughout the network increasing routing overheads. Thus, it is crucial for power control algorithms in a mobile wireless network to adjust and adapt to the dynamic topology and physical environment.;A comparative study of the proposed distributed algorithm with common power strategies in literature illustrates the superior network performance of DISPOW and DDISPOW. The algorithms effectively maintain network connectivity and reduce inter-node interference. They alleviate the dependency of common power strategy on few nodes that are isolated in the network because of their physical location and surrounding network environment. Detail analyses show that DISPOW and DDISPOW are energy efficient and significantly improve the performance of wireless networks.

Type

dissertation

Source

2009_2013.csv

degree

Ph.D.

Program

Engineering

Item