Simulation Framework for QoI Characterization of Sensor Networks in the Presence of Faults

The quality of information delivered by a sensor network depends heavily on the integrity of data produced by the sensor nodes themselves. As such, the ability to model a sensor networking application scenario in the presence of faults that affect sensor data integrity is quite crucial. In this paper we describe a sensor network simulation framework that provides high fidelity modeling of sensor faults, and enables users to study end-to-end quality of information. Our sensor fault simulator is based on the popular, open-source simulator OMNet++ and the associated Castalia package for modeling basic sensor network behaviors. The framework provides the ability to model various elements of a complete application scenario including node deployment, physical process, sensor faults, fault detection, and data fusion. This paper will describe the architecture of the simulation platform, the novel sensor fault model for modeling a variety of fault types and progressive failure behaviors in a parameterized fashion, and the enhancements to Castalia with integrated functionalities for fault detection.