Adapting to Urban Warfare

Abstract

Urban operations currently are of great concern to the defense community. J9, the Experimentation Directorate of USJFCOM, and the Joint Advanced Warfighting Program currently are conducting an experiment to investigate concepts for applying future technologies to joint urban operations. The first phase of the experiment focused on employing future sensors to remotely monitor and understand enemy operations in a foreign city. Characteristics of the urban environment include high building density, a large civilian population, and a cultural environment. These characteristics pose significant challenges for simulation designers. This paper describes the modifications required to adapt the simulations supporting the experiment, JSAF and SLAMEM, to the urban environment. A landscape with a large number of buildings had to be automatically generated and represented in a space efficient manner. Large concentrations of vehicles and pedestrians had to be modeled moving realistically through the city. This behavior had to be automatically generated since it would be impossible to individually control 100,000 entities. Embedding cultural features within the database in the form of building function codes allow civilian entities to automatically plan their movements based on generic daily schedules. Sensor models had to be modified to detect building properties, such as whether a building was fortified. The density of both entities and structures made both movement and intervisibility calculations significantly more expensive, requiring optimization combined with the application of large amounts of hardware. Computation and control was distributed between three CONUS sites and the High Performance Computing Centers at Maui and Wright-Patterson AFB. Limiting and balancing simulation traffic required a major effort. Source squelching was enabled by a distributed data collection system developed to collect data locally on each simulation node while still allowing analysts to perform real-time queries during the experiment.