Providing background vehicles to serve as sensor clutter is an important component for experiments involving sensor detection of clandestine opposing forces. The ClutterSim has been used since 1998 to cheaply simulate large amounts of civilian vehicles for the purpose of confusing or overloading sensors. Clutter vehicles need to be simple enough to be cheap to simulate, but realistic enough that OPFOR vehicles are difficult to detect when they are performing their tasks.

In support of the Urban Resolve 2004 (JUO) experiment, it was necessary to further increase the intelligence of the clutter vehicles to mimic realistic urban traffic patterns, including traveling on the correct side of the road, stopping at traffic lights, parallel parking, stopping for pedestrians, and so forth. This was perhaps the most challenging requirement in the history of clutter development, since it conflicted with one of the primary methods for keeping clutter cheap to simulate: clutter vehicles are completely unaware of all other vehicles, including other clutter vehicles. A further complication was that any solution had to work properly over the network, since clutter is commonly simulated on multiple machines.

This paper describes the solution to this problem: the creation of an network-synchronized traffic intersection controller which permits appropriate behavior of clutter vehicles at an intersection while preserving clutter vehicles' ignorance of the presence of other vehicles and only adding a minimal amount of network traffic.