Skip to content

Latest commit

 

History

History
68 lines (45 loc) · 3.89 KB

README.md

File metadata and controls

68 lines (45 loc) · 3.89 KB
Raw

TRAILS - A Trace-Based Probabilistic Mobility Model

Modeling mobility is a key aspect when simulating different types of networks and a large number of various models has emerged in the last years. They are typically either trace-based, where GPS recordings are re-run in simulation, or synthetic models, which describe mobility with formal methods. Both concepts have advantages and disadvantages. For example, trace-based models are very in flexible in terms of simulation scenarios but have realistic behaviour, while synthetic models are very flexible, but lack realism. To overcome these these issues, a mobility model by the name of TRAILS (TRAce-based ProbabILiStic Mobility Model) has been developed. This model spans the gap between both families and combines their advantages into a single model. This repository contains an implementation of this mobility model in OMNeT++, a C++ based discrete event simulator.

The immediately following section is the procedure for anyone who wishes to get TRAILS up and running after.

TRAILS Setup Procedure

TRAILS is uses the simulation environment of the OMNeT++ simulator and the scripts use Python and Bash. It is only meant for Linux or MacOS environments. The procedure for setting up and running TRAILS is given below.

  • Open a console (terminal)
  • Install and setup OMNeT++ v5.2.1 or higher (use OMNeT++ installation Guide)
  • Install Python 2.7
  • Clone this repository: git clone -b https://github.com/ComNets-Bremen/TRAILS
  • Change directory to TRAILS
  • Download and untar the latest version of OMNeT++ INET from https://inet.omnetpp.org/Download.html
  • Build INET: run ./build-INET.sh
  • Build mobility model testing environment (MobilityModelCheck): ./build-MobilityModelCheck.sh
  • Build TRAILS database and BonnMotion trace file from SFO Taxi Cab traces: run cd ./Traces/SFO-Taxi-Cabs ; ./prepare-traces.sh ; cd ../../
  • Simulate mobility for 4 mobility models (Random Waypoint, SWIM, BonnMotion, TRAILS): run ./run-4-mobility-models.sh
  • Extract results from simulation output files in MobilityModelCheck/results

Creating TRAILS Database from SFO Taxi Cab Traces

Different TRAILS databased can be created with other parameters. Change the parameters given in the script ./prepare-traces.sh in ./Traces/SFO-Taxi-Cabs to suit requirements.

Collecting and Creating Results from Simulation Output Files

There are a set of codes in the log files that gives details about the number of contacts made during a simulations. The following list gives these codes and their meanings.

Code Description
ANS accumulated neighbourhood size
TNRT total neighbourhood reporting times
ACD accumulated contact durations
TNC total number of contacts up to now

Statistics computed from these coded values are as follows.

  • Average Neighbourhood Size = ANS / TNRT
  • Average Contact Duration = ACD / TNC
  • Total Number of Contacts = TNC

Always use the last values listed in the logs. To get the last values, run the following bash commands on a terminal.

grep "TNRT\|ANS\|TNC\|ACD" General-0-20180518-23:16:11-28993-log2.txt | tail

Replace the General-0-20180518-23:16:11-28993-log2.txt file with the log file in in your simulation folder.

A python script has been written to collect the above mentioned results (results-collect.py) and it is available at Scripts/Results-Collect.

Questions or Comments

If you have any questions, comments or suggestions, please write to us using any of the e-mail adresses below.