OPNET Technologies
7255 Woodmont Avenue 
Bethesda, MD 20814

Tel: 240-497-3000
Fax: 240-497-3001
http://www.opnet.com/

OPNET is a registered
trademark of OPNET Technologies
2000 OPNET Technologies 

University:   New Mexico State University
Researcher: Assoc. Prof. John P. Mullen
Department: Center for Stochastic Modeling, Industrial Engineering Department

Research With OPNET: Improving fidelity in the simulation of mobile ad hoc networks in OPNET.

Main focus of this project: At this time, there is a tremendous gap between the performance of Mobile Ad hoc NETwork (MANET) routing protocols predicted by simulation and that actually experienced in the field.  The purpose of this research is to narrow that gap by developing simulation tools that

  1. Better predict actual field performance of MANET protocols,
  2. Allow modeling of techniques presently being developed to manage the unreliability of wireless links,
  3. Not greatly increase the modeling load, and
  4. Do so in a way that has, at most, only a moderate impact on simulation execution time. 

These tools would not only narrow the gap between simulation and field experience, but also allow simulation to be a more effective tool in the design of robust MANET routing protocols and the planning of field experiments.

Approach: The primary function of a MANET protocol is to establish and maintain multihop routes through the net that act as a virtual backbone. Almost all simulations of MANETs assume that fine-grained variations in received signal power do not have a significant effect on performance.  However, field experience shows that the impact of this variation is huge.  Initially, the focus was on efficient models of this fine grain variation.  At this time, the focus has shifted to robust performance analysis and parameter selection, similar to Taguchi's methods.

Results: At this time, we have demonstrated the feasibility and value of our basic approach by comparing performance predicted by our modified models to published results of implementation attempts in test beds and in the field.  However, in order to convert this preliminary work into a useful tool, we need to:

  1. Validate our simulations in more real-world situations,
  2. Incorporate more models of signal variation, and
  3. Provide a better user interface. 

In addition, while we have a method to implement Taguchi and response-surface designs, these methods are not efficient. Thus, we a attempting to develop a more streamlined set of design and analysis tools. Our ultimate objective is the development of effective components which can be used to construct robust MANET Routing Protocols.

Publications from this project:

Related Publications:

Works in Progress:

  • J. Mullen, H. Hwang. "Evaluating the preformance of MANETs under extreme conditions."
  • J. Mullen, X. Hua. "Improved estimates of signal to noise plus interference ratios in wireless ad hoc nets."
  • J. Mullen, Markson Ofoukuo. "Sensitivity of MANET performance to wireless signal characteristics.

This work is supported by a software license grant from OPNET Technogies and other support from the Center fo Stochastic Modeling.

The Center for Stochastic Modeling is a  multi-disciplinary consortium that attempts to develop efficient, realistic models of stochastic systems and use these models to develop effective tools to study and manage such systems.  A current focuses of this group is Mobile Ad Hoc Networks (MANETs).

 

Center for
Stochastic Modeling

Industrial Engineering
201 EC III
New Mexico State University
Las Cruces, NM 88003

 

John P. Mullen

Tel: (505) 646-2958
Fax: (505) 646-2976
email: jomullen@nmsu.edu

       
rev jpm05dec2006