Procedural Modelling of Cities

As the capability of display platforms improves, generating realistic content to realize the potential of these platforms becomes increasingly challenging. Modelers need tools to help them generate realistic content. Tools for the modeling of natural objects such as plants are well-researched, but man-made objects have been largely overlooked. We are working on tools for the generation of realistic cities. Our approach uses agent-based simulation.

Principal Investigators: Ben Watson (NCSU CSC), Uri Wilensky (NU LS/CS) &, Martin Felsen (IIT & UrbanLab).

Other project members: Tom Lechner (NU CS), Seth Tisue (NU LS), Andy Moddrell (IIT), Chris Sexton (NCSU CSC).

Sponsors: NSF ITR award 0326542, Electronic Arts & Maxis.

Thanks to: the Center for Connected Learning, the chil.us project, the Environmental Simulation Center, John Buchanan, Christopher Alexander & Michael Shiffer.

Publications

C. Sexton, B.A. Watson (2007). Vectorization of gridded urban land use data (ref). ACM SIGGRAPH 2007 posters (San Diego, CA).

T. Lechner, B.A. Watson, S. Tisue, U. Wilensky & M. Felsen (2004). Procedural modeling of land use in cities (pdf). Technical report NWU-CS-04-38.

T. Lechner, B.A. Watson, U. Wilensky & M. Felsen (2003). Proceduring city modeling (pdf). 1st Midwestern Graphics Conference (St. Louis, MO).

Videos & Imagery

	Simulated city growth A city develops on a lakeshore. A simple gif animation.
	A large city A fairly mature city. Note areas of gridded and curving roads. Yellow is residential, red is commercial, blue is industrial.
	A SimCity visualization Here we zoom in on a city populated with SimCity's buildings. Note the transition to a curved and meandering layout, which cannot be simulated automatically by SimCity.