Modeling, validation, and simulation of massive self-organizing wireless sensor networks with cross-layer optimization and congestion mitigation techniques

TNO has formulated the ambition of founding a basis for the development of flexible multi-data source and multi-application (ad hoc) sensor networks. These networks are envisioned on a scale that is beyond networks for specific and separate sensor networks. These separate networks need in the future to be combined intelligently in order to assist users with real-time assessment during operations in challenging environments. For this, enabling technologies such as network architectures and model-based data-interpretation are being developed. In this context TNO has developed wireless sensor networks that consist of very small and cheap nodes. Each node can be size of a golf ball and can run for a long time on an A4 battery. They are designed to be deployed on a massive scale (100-10 k or larger). In the tactical environment they are typically lying on the ground. Because it is very difficult to test the performance of such deployments under realistic and controllable conditions, simulation is the only way to get insight in this. This paper focuses on how to study the performance of massive deployments of these sensor networks through simulation. It describes the modeling, validation and simulation of such networks. Features such as antenna patterns, radio propagation/path loss, terrain, medium access and transport protocols, routing, and simulation parallelization are discussed. Model validation experiments for the antenna pattern, the path loss propagation and communication performance for nodes lying on the ground will be discussed. It further describes the implementation of a promising energy aware MAC and transport protocol that are proposed in the literature and are suitable for congestion mitigation and cross-layer optimization. Results on timeliness of messages, bandwidth and power consumption are discussed.