仿真DTN网络，通常需要产生节点移动的网络拓扑。目前，ns3支持两种移动场景产生器(Mobility scenario generators)，即ns-2 setdest和​​BonnMotion。本文结合实例讲解​​BonnMotion使用方法，即先根据相应的模型产生节点移动数据，再根据相应的平台将这些数据转化为符合目标平台的格式。

1. 安装

官网下载BonnMotion安装包，解压，进入目录运行./install即可(需要事先安装jdk，sudo apt-get install default-jdk)。官方使用指南点这里进行下载。

2. 产生移动场景数据

命令./bin/bm -hm查看可生成的场景，如下：

Available models:
Boundless               construct Boundless mobility scenarios
ChainScenario           links different scenarios
Column                  create movement scenarios according to the Column model
DisasterArea            create extended catastrophe scenarios according to the Disaster Area model
GaussMarkov             construct GaussMarkov mobility scenarios
ManhattanGrid           construct ManhattanGrid mobility scenarios
MSLAW                   construct MSLAW mobility scenarios
Nomadic                 create movement scenarios according to the Nomadic community Mobility model
OriginalGaussMarkov     construct mobility scenarios according to the original Gauss-Markov model
ProbRandomWalk          construct Probabilistic Random Walk mobility scenarios
Pursue                  create movement scenarios according to the Pursue Mobility model
RandomDirection         construct Random Direction mobility scenarios
RandomStreet            construct RandomStreet mobility scenarios
RandomWalk              construct RandomWalk mobility scenarios
RandomWaypoint          construct RandomWaypoint mobility scenarios
RandomWaypoint3D        construct RandomWaypoint (3D) mobility scenarios
RPGM                    create movement scenarios according to the Reference Point Group Mobility model
SLAW                    construct mobility scenarios according to the Self-similar Least Action Walk model
SMOOTH                  construct SMOOTH mobility scenarios
Static                  construct static scenarios with no movement at all
StaticDrift             construct static scenarios with a drift
SteadyStateRandomWaypoint  construct Steady State Random Waypoint mobility scenarios
TIMM                    construct Tactical Indoor Mobility Model mobility scenarios

本例产生RandomWaypoint移动模型，random waypoint model(方向与速度都随机选择)是随机模型的一种，节点在仿真区域(x*y)内随机选择一个方向，在最小与最大速度之间随机选择一个速度，节点移动到目的地后，在0与最大停止时间随机选择一个时间停止不动^[1]。命令./bin/bm –hm RandomWaypoint查看使用方法，如下：

Scenario:
    -a <attractor parameters (if applicable for model)>
    -c [use circular shape (if applicable for model)]
    -d <scenario duration>
    -i <number of seconds to skip>
    -n <number of nodes>
    -x <width of simulation area>
    -y <height of simulation area>
    -R <random seed>
RandomSpeedBase:
    -h <max. speed>
    -l <min. speed>
    -p <max. pause time>
RandomWaypoint:
    -o <dimension: 1: x only, 2: x or y, 3: x and y>

这里，只有节点数-n是必须，其他都有默认值，运行命令./bin/bm -f test_scenario RandomWaypoint -n 100产生两个文件test_scenario.params和test_scenario.movements.gz(尽管可以用gzip -d filename.gz解压，但没必要，直接vim就可以打开，而且后续转换格式需要.gz这个文件)。test_scenario.params存放的是参数，从这里可以看到参数的默认值，如下：

model=RandomWaypoint
ignore=3600.0
randomSeed=1398701966786
x=200.0
y=200.0
duration=600.0
nn=100
circular=false
dim=3
minspeed=0.5
maxspeed=1.5
maxpause=60.0

test_scenario.movements.gz是BonnMotion规定的数据存放格式，下一步将其转换成ns3可以使用的格式。

3. 转化成ns3使用的格式

BonnMotion可以将上述产生的场景数据转化成符合不同仿真平台的数据格式(如ns2, ns3, Glomosim, The ONE,)，用命令./bin/bm -ha查看所支持的转换，摘抄部分如下：

Available apps:
Cut                 saves a certain timeframe from one scenario into a new file
CSVFile             creates a simple CSV file for processing in 3rd-party tools
Dwelltime           analyses scenarios according to Bettstetter
GlomoFile           creates a movement file for Glomosim (2.0.3) and Qualnet (3.5.1)
GPXImport           converts GPX files to Bonnmotion format
IntervalFormat      converts scenario file into interval format
LinkDump            dumps informations about the links
LongestLink         Longest Link Metric
NSFile              creates movement files for ns-2
ScenarioConverter   converts scenario types
SPPXml              converts scenarios according to Horst Hellbruecks XML schema
Statistics          calculates various statistics for movement scenarios
TheONEFile          converts scenarios to the ONE file format
Visplot             visualises node movements
WiseML              converts scenario files to WiseML format

用命令./bin/bm -ha <application name>进一步查看特定应用的使用方法，尽管BonnMotion没有专门为ns3提供支持，但ns3可以通过Ns2MobilityHelper类读取ns2的移动文件(movement files)。命令./bin/bm -ha NSFile查看使用方法，如下：

./bin/bm NSFile [-b double] [-d] -f <scenario>
-b double  增加边缘(margin)，默认值为10
-d  禁用模块特殊转换(如灾区disaster area)，使用标准转换
-f <scenario>  上述生成的场景文件，这里不需要输入后缀名.params或.movements.gz

在本例，运行命令./bin/bm NSFile -f test_scenario，生成NSFile格式的移动文件test_scenario.ns_movements 和test_scenario.params，test_scenario.params包含一些参数，如下：

set val(x) 220.0
set val(y) 220.0
set val(nn) 100
set val(duration) 600.0

文件test_scenario.ns_movements由一系列节点移动情况组成，取第一条如下：

$node_(0) set X_ 126.37346418288988
$node_(0) set Y_ 135.94261044575848
$ns_ at 0.0 "$node_(0) setdest 96.77496900599616 177.38003952427124 1.246967019035919"
$ns_ at 45.435969720538196 "$node_(0) setdest 107.85949715020739 98.22059349334985 0.7091627087954391"
$ns_ at 196.16204414948334 "$node_(0) setdest 73.94845981139225 206.62194932677775 0.815787004338213"
$ns_ at 393.6431734580033 "$node_(0) setdest 103.07672579116219 46.015637307213 1.2760546622517224"
$ns_ at 533.0677403031077 "$node_(0) setdest 176.96853459134928 104.20848371058558 1.405231787345907"

前两行是节点的初始位置，其后的每一行描述每一次移动，抽象出来，如下：

$node set X_ \<x1\>
$node set Y_ \<y1\>
$ns at $time $node setdest \<x2\> \<y2\> \<speed\>

上述的语句含义是，在时间点time，节点node以速度speed从(x1, y1)移动到(x2, y2)。

4. 导入ns3

参照例子src/mobility/examples/bonnmotion-ns2-example.cc，将上述场景数据导入ns3，主要代码如下：

CommandLine cmd;
cmd.Parse(argc, argv);

NodeContainer nodes;
nodes.Create(100);

std::string traceFile = "scratch/test_scenario.ns_movements";
Ns2MobilityHelper ns2 = Ns2MobilityHelper (traceFile);
ns2.Install ();

Ns2MobilityHelper类只有一个函数，所以重点在于trace file。运行命令 ./waf –run scratch/test_scenario –vis，点击仿真，观察节点的移动，如下：

图1 BonnMotion导入ns3例子

让我感到纳闷的是，100个节点，为什么只扫描到99个节点？如下：

scanning topology: 100 nodes…
scan topology… 99 nodes visited (99.0%)
scanning topology: calling graphviz layout
scanning topology: all done.

不足的是，BonnMotion支持的随机模型几乎是完全随机的，即服从泊松分布，而现实的网络几乎都是服从幂律分布，如何产生服务幂律分布的网络？如果您有好的方法，希望您能跟我分享，Jelline@126.com

参考资料：
[1]维基百科词条：Random waypoint model

附：BonnMotion简介

BonnMotion is a Java software which creates and analyzes mobility scenarios and is most commonly used as a tool for the investigation of mobile ad hoc network characteristics. The scenarios can also be exported for several network simulators, such as ns-2, ns-3, Glo-MoSim/QualNet, COOJA, MiXiM, and ONE. BonnMotion is being jointly developed by the Communication Systems group at the University of Bonn, Germany, the Toilers group at the Colorado School of Mines, Golden, CO, USA, and the Distributed Systems group at the University of Osnabr ̈ck, Germany.

赞赏

微信赞赏支付宝赞赏