A PDF file should load here. If you do not see its contents
the file may be temporarily unavailable at the journal website
or you do not have a PDF plug-in installed and enabled in your browser.
Alternatively, you can download the file locally and open with any standalone PDF reader:
http://downloads.hindawi.com/journals/sp/2004/921065.pdf
A Mobility and Traffic Generation Framework for Modeling and Simulating Ad Hoc Communication Networks
Scientific Programming
1058-9244
A mobility and traffic generation framework for modeling and simulating ad hoc communication networks1
Chris Barrett 2
Martin Drozda 1
Madhav V. Marathe 2
S.S. Ravi
James P. Smith 2
Basic
Applied Simulation Science (CCS-
) Los Alamos National Laboratory
P.O. Box
Los Alamos
USA E-mail:
barrett
marathe
jpsmith}@lanl.gov
0 The work is supported by the Department of Energy under Contract W-7405-ENG-36
1 University of Hamover, FG Simulation and Modellierung , Welfengarton 1, 30167 Hannover, Germang
2 grant from Los Alamos National Laboratory and by NSF Grant CCR-97-34936
We present a generic mobility and traffic generation framework that can be incorporated into a tool for modeling and simulating large scale ad hoc networks. Three components of this framework, namely a mobility data generator (MDG), a graph structure generator (GSG) and an occlusion modification tool (OMT) allow a variety of mobility models to be incorporated into the tool. The MDG module generates positions of transceivers at specified time instants. The GSG module constructs the graph corresponding to the ad hoc network from the mobility data provided by MDG. The OMT module modifies the connectivity of the graph produced by GSG to allow for occlusion effects. With two other modules, namely an activity data generator (ADG) which generates packet transmission activities for transceivers and a packet activity simulator (PAS) which simulates the movement and interaction of packets among the transceivers, the framework allows the modeling and simulation of ad hoc communication networks. The design of the framework allows a user to incorporate various realistic parameters crucial in the simulation. We illustrate the utility of our framework through a comparative study of three mobility models. Two of these are synthetic models (random waypoint and exponentially correlated mobility) proposed in the literature. The third model is based on an urban population mobility modeling tool (TRANSIMS) developed at the Los Alamos National Laboratory. This tool is capable of providing comprehensive information about the demographics, mobility and interactions of members of a large urban population. A comparison of these models is carried out by computing a variety of parameters associated with the graph structures generated by the models. There has recently been interest in the structural properties of graphs that arise in real world systems. We examine two aspects of this for the graphs created by the mobility models: change associated with power control (range of transceivers) and variation in time as transceivers move in space.
Mobile and ad hoc networks; mobility models; simulation and modeling; graph theory
1. Introduction
1.1. Simulation of ad hoc networks
An ad hoc wireless mobile network is a collection of
mobile transceivers that communicate via radio
transmission. There is no wireline network to support the
movement of packets. The communication network is
formed spontaneously when transceivers activate their
radios. Moreover, movement of the transceivers
constantly changes the connectivity of the network. These
networks are gaining popularity due to their
applicability in a number of important situations such as
battlefield communications, emergency management and
response, etc. The topic of designing and building large
scale ad hoc networks is an area of active research.
The dynamic connectivity of the network that is
induced by the constant movement of transceivers implies
that the performance of the communication network
depends crucially on how the underlying transceivers
move. Indeed, as has been observed in the literature
(see for example [
9,10,21,26,27,38
]), parameters such
as the velocity of transceivers, their mobility patterns
and their spatial density significantly affect the network
performance.
The goal of the AdHopNet project at the Los Alamos
National Laboratory (LANL) is to develop a scalable
simulation-based tool that can be used for the design
and analysis of ad hoc communication networks being
built as a part of SUO-SAS project at DARPA [
8
] and
the Urban Infrastructure Suite of analytical tools being
developed as a part of the National Infrastructure
Simulation and Analysis Capability (NISAC). The focus of
this paper is on some modules of this simulation-based
analysis tool. These modules produce mobility
patterns, and the data generated by these modules in
conjunction with a module that generates communication
activities is used to drive the packet movement
simulator of AdHopNet. The modules for generating
mobility data were developed originally for AdHopNet.
However, they can be used in conjunction with other
simulation tools such as GloMoSim and ns-2 [
3,4
]. The
flexible framework within which the modules operate
allows us to create both synthetic and realistic mobility
patterns along with appropriate communication
activities (or calling patterns). Although our framework is
des (...truncated)