Designers predict how people will use the spaces they build, and those predictions are almost never validated against observed behavior. The GAINS team develops the SPACE Engine, the simulation half of the SPACE Lab research program, which generates synthetic populations of behaviorally realistic NPC agents and deploys them in virtual analogs of real sites. The simulation's output matches the format of the SCOUT observation system, enabling direct GIS- and parametric software-based comparison between simulated and observed pedestrian behavior.

Working in Unreal Engine 5, this team develops the SPACE Engine through a sprint-based agile process. Sub-teams of two to three students own individual system components: NPC agent architecture, per-agent behavioral and physiological variable systems, environment evaluation and scoring logic, population spawn management, and output pipelines matched to the format of field observation data. Every behavioral variable in the system requires a published literature source. The design is grounded in environmental psychology research, not game design conventions.

VIP GAINS operates through the Purdue Vertically Integrated Projects program. Students join for multiple semesters, building institutional knowledge across cohorts, with returning members taking on component leadership roles and helping onboard incoming students each term. No prior Unreal Engine experience is required. Students who contribute to architectural decisions, documented in sprint records and version control history, are eligible for contribution credit and co-authorship on peer-reviewed publications.