Low intensity conflict often involves heavy interaction with civilian populations during activities ranging from patrols to convoy operations. As a result, these interactions are important to realistic training. This paper discusses applied research into representations of civilian traffic rules in Army computer generated forces (CGF) applications. While this work emphasizes the needs of Army driver trainers as a primary focus, proper CGF representation of civilian traffic rules are critical to many other training use cases, including convoys, checkpoints, first responder training, and recognition of aberrant behavior in civilians that may be a precursor to an insurgent attack.

This research investigates existing capabilities, deficiencies, and possible future needs of Army entity-level simulations. Two paths were investigated in parallel. One path is more practical and applied, recognizing the current limitations of modeling & simulation source data and system architectures. This path provides functional improvements for near-term transition to programs. In parallel, a more advanced effort is investigating solutions that are more complex and sweeping, therefore complicating practical application. However, this thread illustrates what is possible as a long term solution. This paper discusses this dichotomy of priorities and how these differing goals were addressed.

There are a number of challenges with implementing civilian traffic functionality in multiple phases of modeling and simulation development. Common geospatial sources lack fundamental data needed for accurate traffic simulation, and significant deficiencies in entity-level CGF systems complicate more advanced behaviors. New concepts for path planning and entity avoidance are needed based upon common traffic rules and patterns such as "lane awareness". This paper explores the challenges in current M&S technologies, describes implemented and proposed solutions, and describes how lessons-learned can be applied beyond traffic simulation into other pattern-of-life simulation use cases.