Future Now
The IFTF Blog
Simulation in Pedestrian Movement Modeling
I recently returned from the 3rd annual meeting on Information Systems for Crisis Response and Management where I gave a paper I wrote on how pervasive sensor networks will support sociological studies of future disasters. I was fascinated to find how much progress is being made in pedestrian movement modelling - researchers are really trying to understand how people, the built environment, and hazards interact during disasters to improve egress from buildings, plazas, stadiums and the like. There is even something of a cottage industry springing up in the UK, where there is a long history of securing public space in the IRA- and Al Qaeda eta. Crowd Dynamics and Intelligent Space (started by my colleague Andy Smith) both offer agent-based pedestrian modelling consulting services.
I think this is an interesting tipping point that could have huge implications for urban design. Instead of over-engineering structures to sustain heavy damage so they can stay up long enough to evacuate, we might design them more like automobile frames - which are designed to destroy themselves to protect the occupants. So, for instance the World Trade Center lasted for about 2 hours - if pedestrian modeling had been used for egress analysis of that lightweight structure, more people might have survived.
There are hints that simulation of entire building ecosystems are going to be the wave of the future. My friend Paul Seletsky, Digital Design Director at Skidmore, Owings, Merrill in New York is trying very hard to get that firm to start simulating buildings from the first stage of design - they see it as a way of expanding the architect's role and involvement with projects well after the traditional point where they hand over the plans to the construction contractor. Also, MIT's City Development and Design program has been experimenting with design tools such as the Luminous Table and Luminous Clay (image below) that allow for augmented reality-type simulations of wind flow and glare... adding pedestrian or traffic simulations would be trivial once that framework is established.