| Sensors used in protective applications are typically placed on perimeters or over areas in an evenly distributed pattern. However, such patterns may actually be sub-optimal, since environmental factors may make some forms of attack more or less likely than others. We describe a protective application of sensors for detecting underwater threats in an urban estuary environment. We demonstrate that environmental information, derived from a computational river current model, can be utilized to optimize sensor placement, increasing detection rates and decreasing the number of required sensors. Simulation results show a significant improvement in detection for a given number of sensors; alternatively, fewer sensors can be used while still maintaining the detection rate of a conventional approach. |