Fast simulation methods to predict wireless sensor network performance

With the increasing capabilities of Wireless Sensor Networks (WSN), complexity and expectation of the WSN applications increase as well. In order to make design-space exploration possi-ble, it is necessary to have fast models that provide adequate in-sight in system behavior. In this paper, we propose a highly ab-stracted, hierarchical, system-level modeling method for WSN. Based on the model properties, fast simulation techniques can be applied. First, an abstract discrete event simulation based on a Probabilistic Graph Model (PGM) is introduced. Then, a fast Monte Carlo simulation approach is proposed for speeding up the simulation process. This approach combines Stochastic-Variable Graph Models (SVGM), providing a high level of abstraction, with shortest path calculations. As a case study, a temperature mapping application in a gossip-based WSN is used, showing a good accuracy of the model predictions