Bridging the Gap Between Citizens and City Decision-Makers
- By Jennifer Jones
- Sep 30, 1997
During a computer-simulated drive over a new design for San Francisco's Bay Bridge, elected city officials gasped as they accelerated over an incline into a panoramic view of-blue sky. Created by a potent new combination of software tools and visualization techniques, the demonstration provided a startling re-creation of what it would feel like to drive over a new "skyway" design of the bridge. And it felt frighteningly like a lift-off.
To the group, which included California Senate president Bill Lockyer, the "urban simulation" demonstration helped show what any number of plastic models could not: a design flaw-discernible only at certain speeds-that could be unpopular with motorists and, therefore, damaging to any politician who approved the $1.3 billion design.
The demonstration was the centerpiece of a California Transportation Department (Caltrans) campaign to help citizens swallow the bitter pill of higher taxes and increased tolls, which were needed to finance a replacement of the earthquake-weakened bridge. It also was the major public debut of a technology that many believe will be an essential decision-support tool in coming years as city planners try to balance an increasing number of community interests.
"The public demands-and is getting-a larger voice in the planning decisions of publicly funded construction projects through the use of urban-simulation technology," Lockyer said. "Experts predict that within three years, every major metropolitan area in the world will be using urban simulation as a means to build consensus between the people, industry and government."
Ironically, the Bay Bridge fly-through helped convince stakeholders that a more expensive option-a "landmark" design reminiscent of the current Bay Bridge-was preferable to the simpler, cheaper, "viaduct" design that California Gov. Pete Wilson wanted to build. "The city's decision-makers were all leaning toward the viaduct option mainly because of cost," said Matthew Cuerdon, the architecture, engineering and construction account manager for Silicon Graphics Inc., which sponsored the simulation.
But not only did the viaduct design turn spooky on the virtual drive, it lacked the architectural flair of competing designs. "During the simulation, everyone noticed that the towers projected above the horizon, while the viaduct was unidentifiable as a bridge structure," Cuerdon said. "Everyone said the towers were impressive aesthetically and distinguished the bridge from all of the background clutter. All were willing to say it was worth the extra $200 million."
The Bay Bridge demonstration is the latest in a series of projects nationwide that promise to bring urban simulation out of the arcades and into the real world of urban planning and engineering. In Los Angeles, Philadelphia and Tucson, Ariz., teams are at work creating real-time, 3-D software models of entire urban ecosystems for applications that include transportation planning, site management, emergency response and historic restoration.
Caltrans plans to launch its own urban-simulation project in the next six months, having pegged endless uses for the technology. "We have a lot of public meetings and hearings where we have to show our projects," said Caltrans public affairs chief Greg Bayol. "Most of our projects are big, and they change peoples' environments. For example, we are building a lot of sound walls here, and we really have a difficult time showing people what a sound wall will look like and how it will change the space around their homes."
The projects have been spurred by three trends: lower prices for high-powered workstations; leaps in computing power that speed the processing of high-resolution images and high-density textures; and new software techniques that integrate a variety of data formats, including geographic information system data, computer-aided design (CAD) drawings, aerial photos, video clips and computer simulations.
To create the Bay Bridge simulation, for example, engineers used Coryphaeus Inc.'s MetroSim, a real-time 3-D application running on an SGI Onyx2 high-end graphics system. Three-dimensional models of each design were combined with real-world data, including Bay area terrain data from the U.S. Geological Survey, satellite imagery of the bay and high-resolution models of nearby landmarks.
Urban-simulation technology borrows from several fields: military simulation, computer gaming, 3-D graphics, animation and CAD. "The easiest way to describe it is to mention flight-simulation trainers; it works the same way," said Bill Jepson, director of computing at the Department of Architecture and Urban Design at the University of California, Los Angeles. "Using a mouse, you can go through different models and look at things from any and all perspectives."
Jepson, who coined the term "urban simulation," is building "Virtual L.A," a project to map the entire city of Los Angeles in 3-D. "Our objective is to have the entire L.A. basin, which is 4,000 square miles," he said. "We now have 15 to 20 blocks done to the accuracy at which you can read the graffiti on the walls and see the signs in the windows. We want to have an entire data set, which will include upwards of a terabyte of data on one or more servers, that our clients will be able to use to roam through the streets of L.A."
Four years ago, Los Angeles awarded Jepson a $175,000 grant to buy an SGI Onyx2 computer, which became a turning point in his 15 years of work with the technology. "That allowed us to begin working with real-time texture mapping," Jepson said. "For the first few years, we didn't have the ability to do textured mapping or photographic imagery with models. Now it is no longer a dream. It's working, and we're doing real projects."
Indeed, advances in texture memory-which enable urban simulation to include even the smallest details of a scene-have helped spark the technology. In the past, only about 4M of texture memory was available on most computers. In contrast, today's SGI O2 and Onyx2 lines provide for hundreds of megabytes of texture memory.
That is important for applications such as the Bay Bridge, where texturing is necessary to simulate the sky, steel cables and other minute objects necessary to dupe the user into a sense of real-time travel. "What brings life to a scene is the detailed textures on most or all objects you see," said Kempton Izuno, manager of Asia-Pacific operations for Coryphaeus, who compared the process to pasting on wallpaper.
The breakthrough in texturing capabilities in turn sparked advances in the software required to generate and run 3-D simulations. Two companies now dominating the market for such software are Coryphaeus and MultiGen Inc., a San Jose, Calif.-based company that got its start in the field by developing gaming applications. MultiGen makes Creator, a 3-D modeling package for Unix and Microsoft Corp.'s Windows NT, and GO, a program to view them.
In addition to software tools, high-end urban simulation requires scalable, multiple-processor-powered hardware capable of supporting real-time interactivity.
"If you go to the cinema, they are projecting the film at 24 frames per second; it's about 30 frames per second for TV," said Drew Henry, SGI's director of advanced graphics marketing. "The perfect [speed] for simulation applications is 60 frames per second. The computer also has to be capable of projecting and calculating at the same time."
But not only is urban simulation technology-intensive, it can be labor-intensive as well. For example, Virtual L.A. is being built largely by a team of UCLA students, 12 of whom spend a combined 20 to 40 hours per city block taking photographs with a digital camera to generate the level of detail contained in the application.
Jepson listed the steps necessary to build his virtual city. "You start with information that you can typically get from a city," including coordinates, accurate street addresses and an accurate linear map of the city streets, he said. Superimposed on that are aerial photographs, which "give us [building] footprints or roof prints and trees and landscaping." Using modeling programs-in this case, MultiGen software-the image then can be "pulled up" into three dimensions.
The next layer of simulation is added by the students. "Team members walk up and down the street photographing every facade of every building with a digital camera," Jepson said. "We then paste up a photograph of a building, making the photo fit the aerial footprint.
"The image requirements for doing something like this are so stringent," he added. "Compare this with those 3-D games that are so popular right now. What makes this different is that those games use the same image a million times. We can't do that. We have to be able to have members of the community come in and recognize something that they have been looking at their entire lives, possibly their own homes."
Jepson estimates Virtual L.A. will cost $3,000 per block to construct. "Once you have a 3-D model of a city, you can begin building on it on a piece-by-piece basis, and it fits together," he said. "Right now we have a 10- to 15-block size of it done, and that model is very easy to update."
Although there is no formal arrangement between the city and Jepson's team concerning access to the application, it is generally available to city officials. In one case, the Virtual L.A. model was used to demonstrate how an ordinance requiring tree planing might affect Pico-Union, a poorer section of the city.
"We used urban simulation to show what it would have looked like if this had been done 20 years ago," Jepson said. "We planted the trees and pushed a slider to watch them grow. It was quite interesting. But when we brought the community in to show them, there was a lot of grumbling. The problem here is that Pico-Union is not a nice area, and drug dealers could hide in those trees and jump down and mug people as they passed by. So we simply changed the trees from lush deciduous trees to palm trees."
Virtual L.A. has caught the eye of city planners in Tucson, Ariz., who also want to use the technology as consensus-builder. The city had started to develop Transview, a World Wide Web site to help coordinate data on road closures and construction projects. "We wanted to enhance public cooperation and reduce problems like utilities cutting open newly paved roads," said Chris Wolff, system administrator for the city's Transportation Department pavement management system. "But we've decided to take that a step further after seeing Virtual L.A.
"We're planning on a digital terrain model we can fly," Wolff added. "It will incorporate aerial orthophotography to a one-foot pixel level of detail. We will feed that into a simulation and start building roadway projects and other transportation facilities."
Giving city decision-makers and the public a look at what they do not understand or do not typically consider in making planning decisions is one of the goals of Philadelphia, another candidate for urban simulation. ModelCity Philadelphia, a prototype simulation of the city, is being developed by Bentley Systems Inc., an Exton, Pa.-based software company.
"There is a disconnect in engineering data and publicly available [Web] browsers and client software," said Yoav Etiel, Bentley's senior vice president of marketing and a former town planner. "This new technology is needed to bridge that gap. Bentley introduced the product line so that engineers and town planners can use any browser to access information typically stored in engineering databases."
In Philadelphia, the technology may be as useful below the streets as it is above. "The underworld of a street is something that is not recognized by the average citizen until they come upon a project and wonder why it is taking so long," said Larry Moy, Philadelphia's commissioner of streets.
And the same goes for commercial developers, he added. "They want a timely completion of projects, but if they are not aware of the vast infrastructure of utilities, they are not able to plan properly."
ModelCity Philadelphia is being positioned as a showpiece for other metropolitan areas, Etiel said. "Urban-planning decisions have a major impact on the lives of many people," he said. "If a new road is being built, everybody is an expert on where it should go and where the exits should be. For a politician, it would be a nightmare to make the wrong decision-a decision based on professional advice not consistent with the public will."
To get more in touch with their public, Tucson's transportation officials intend to haul computers to neighborhood meetings to showcase the technology. "We will soon have the visuals to display what a project will look like, and we can have the public immerse themselves in that simulation," said Wolff, who believes urban simulation is part of a "paradigm shift" in how cities sell projects to the public. "Before, the public would have to come to the city to see what projects [would] look like," he said. "Now we are going to them."
And really, that is the whole point of urban simulation. "It adds to the consensus-building between all of the stakeholders in a specific project prior to the commitment of funding, which is often the most lengthy part of a project," SGI's Henry said. "It can go a long way in terms of getting people to agree, be they the general population, political leaders, architects, financial institutions or local building departments."
"God, we're slow learners," said Will Orr, director of the advanced technology program for Scottsdale, Ariz. "We have finally reached the realization that we need more and more data about our community to manage the community wisely and to cope with what we realize now is very rapid change."
Scottsdale has been using remote sensing, modeling and simulation technology, and it is looking to take that to the next level: urban simulation. "It can improve the defensibility of a decision, and any senior person, be they elected or otherwise, wants their decisions to be defensible," Orr said.
"My sense from elected officials is that there are very few of them that don't want to do well for their community," he said. "With urban simulation, there is the promise of some tools that will let them make better decisions and improve their long-term vitality."
is a staff writer for civic.com. She can be reached at email@example.com.
* * * * *
Los Angeles Hones Virtual L.A. Applications
Bill Jepson, who is considered the father of urban simulation, is building "Virtual L.A," a project to map the entire city of Los Angeles. "There are just a million applications for a system like this," he said. The applications Jepson and his team are pursuing include:
* Consensus-building. "Los Angeles has, at best, a spotty record when it comes to new development being sensitive to the needs of the surrounding community," wrote Jepson and Scott Friedman, a University of California, Los Angeles, urban-simulation team member, in a paper on the project. "We have found that designers, architects, developers and consultants are able to identify real problems and remedy those problems long before the first hole on a new development is dug."
* 3-D navigation systems. The UCLA team is in the process of creating a 3-D car-navigation system that will display "an out-the-window view from the vehicle or fly up and track the vehicle from afar."
* Emergency response management. ER commanders will be able to locate vehicles and technicians in real time to provide more effective command and control instructions using visual cues from the Virtual L.A. database.
* 3-D information systems. Jepson plans to use the system not only to plan and design a new hospital at UCLA but to manage it as well. The team has talked to the contractor about linking original engineering drawings of the building to the walls rendered in the model. Then engineers could click on a wall "and pull up the appropriate as-built drawings."
* Historic reconstruction. The UCLA team is creating a model of ancient Rome. "Because we have incorporated the concept of time (other than real time) into the system, we are creating a history of Rome by adding attributes to graphic objects...specifying creation and dissolution dates," Jepson and Friedman wrote. "When loaded, the system recognizes the time-dependent nature of the databases and adds a time slider to the simulation controls. The user can then interactively select...a time or date...to begin the simulation."