Suppose you are a top U.S. air defense commander boarding on the E-3 Sentry AWACS aircraft, who is in charge of making a strategic decision on the intrusion of enemy forces in a timely manner. You would be provided with tens of thousands of tactical data arriving from every corner of world every minute. Your job is to quickly grasp, and react to, changes in the evolving battle; you must possess good situation awareness(SA). This problem is known as command and control in the military community. However, understanding the huge positional data and combining it into a single comprehensive view of the battlefield can be extremely difficult, error prone and time consuming. Thus we need to reduce the information complexity.
We postulate that geometric queries on the positional data set and their
appropriate visual presentation can reduce such a cognitive overhead. In
this dissertation, we present various techniques from visualization and
computational geometry to tackle those problems. The principles of human
visual perception play an important role in this work too.
We identify various visual tools that help a military commander to achieve
situation awareness of the battlefield,
and provide techniques to efficiently compute them. Such visual tools are integrated into a battlefield visualization package, see Figure 2. The package also includes terrain visualization which has been traditionally favored in the battlefield visualization.
Computationally, we explore the techniques from the point of view of dynamic computational geometry. In particular, Delaunay triangulation and its dynamic update have been extensively investigated. We also consider how to visualize the result in both computationally and perceptually efficient ways.
Finally, using behavioral animation techniques, we simulate the battlefield
which provides the inputs to the visual computation.