Comparing a Real-Life WSN Platform Small Network and its OPNET Modeler Model using Hypothesis Testing

Gilbert E. Pérez, Ivica Kostanic

To avoid the high cost and arduous effort usually associated with field analysis of Wireless Sensor Network (WSN), Modeling and Simulation (M&S) is used to predict the behavior and performance of the network. However, the simulation models utilized to imitate real life networks are often used for general purpose. Therefore, they are less likely to provide accurate predictions for different real life networks. In this paper, a comparison methodology based on hypothesis testing is proposed to evaluate and compare simulation output versus real-life network measurements. Performance related parameters such as traffic generation rates and goodput rates for a small WSN are considered. To execute the comparison methodology, a “Comparison Tool”, composed of MATLAB scripts is developed and used. The comparison tool demonstrates the need for model verification and the analysis of good agreements between the simulation and empirical measurements.

Full Text