Modeling emergency events to evaluate the performance of time-critical WSNs

Wireless sensor networks designed for critical tasks must be able to offer near real-time feedback about emergency events, allowing to supervise these events with a reasonable degree of accuracy within strict delay bounds. Achieving such high responsiveness in a distributed environment requires several improvements to currently available WSN protocols and technologies at different layers; such enhancements typically rely on simulation, at least in a preliminary phase. To accurately evaluate the performance of WSNs in near real-time event tracking we developed an event generation framework compatible with the ns- 2 simulator that includes tools to model intruder detection events, as well as fire and gas propagation scenarios. In this paper we describe the analytical models developed, and then we present the potential capabilities of the proposed framework, along with some visual examples of different types of events which confirm its correct behavior.