Situation Awareness: A Network Centric Approach

(ABSTRACT) Situation (al) awareness (SA) is critical to analyze, predict and perform tasks effectively in a dynamic environment. Many studies on SA have ignored network dynamism and its effect on SA, focusing on simple environments. Many studies involving the network and SA have refrained from attempting to model information space dynamism (i.e. dynamic scenarios which may have more than one probable outcome). Few studies have identified the need for a flexible, robust and overarching framework which could model both the network and information space dynamisms and provide for analysis of different types of networks (heterogeneous/homogeneous) at multiple scales. We utilize the NCOPP (Network Centric Operations Performance & Prediction), a uniform framework with " plug-&-play " capabilities to provide analysis and performance prediction of networked information systems. In this work, we demonstrate the flexibility of the NCOPP framework and its ability to model a hierarchical sensor system satisfactorily. We model the network & information space dynamisms using probability and statistics theory (e.g. Bayesian prediction, probability distribution curves). We model the behavior of entities/nodes involved in the process of sharing information to achieve greatly improved situation awareness about a dynamic environment within hierarchical information network systems. Our behavior model mathematically represents how successful/unsuccessful predictions critically impact the achievement of effective situation awareness. In the behavior model, we tie together the cost of considering predictions which accounts for limited resources and the indirect effect of unsuccessful predictions. We research and show how the NCOPP framework can model real world networked information systems at different levels of granularity. We leverage the framework's capabilities to perform experiments that not only assist in an objective comparison of distributed information filtering and central data processing paradigms but also provide important insights into the effect of network dynamism on the quality and completeness of information in the system. We demonstrate the ability of incorporating key network information, in the process of achieving SA to improve the performance of the system. We exhibit the improvement in the performance achieved with inclusion of the network characteristics during dynamic allocation of resources. We were able to show that simple hierarchical filtering (via distributed processing) results in significant reduction in the information with regards to " false alarms " when compared to systems employing central information processing. Experimental results show a direct positive impact in the completeness of SA when information sharing in hierarchical systems is supplemented by network delay information. Overall, …