Being There

Automotive Manufacturing & Production, Dec, 2000 by Gary S. Vasilash

When it comes to computers for design and engineering, General Motors probably leads the auto world [given its size, this speculation is not particularly risky]. Its silicon, software and smarts are being put to good use as its designers and engineers are now moving from the 2D images on screens and clay models to the seemingly 3D world of virtual reality.

Objects have their benefits. But they also have limitations. But what to do when you are designing objects yet want to minimize your limitations. To say that the auto industry has embraced computer-aided design (CAD) is to make about as bold a claim as to say that moviegoers have embraced popcorn. Regardless of which of the world's OEMs it is (there is no geographic limit), regardless of how they describe their digital design domain (many companies tend to be as imaginative in this regard as they are in their vehicle platforms), the car makers employ CAD because they all believe in the power of advanced mathematics to make better vehicles more efficiently...but then there is the utilization of material that has been used for at least as long as homo has been sapient: clay.

Designers make clay models of their vehicle designs. Sometimes the models are full size. Sometimes there is a scale model (often one third size).

The scientists and engineers have developed custom software called "VisualEyes." They are now at version 4.0. As Smith points out, "This is not research code--it's not shaky. It doesn't break all the time." They've worked exceedingly hard at making this software capable of providing images that are high resolution, stereoscopic and full-sized. This is virtual reality (VR). it is kicking up CAD to a whole different level. Way up. Larry Peruski, who works with Smith, said that designers "need a full-size model to use their intuitive sense of depth and curvature, If you put a six-inch cube up there [on the screen], you ought to be able to walk right up to it and make sure it's six inches." The images appear to be real. As Smith remarks, "You can't make decisions based on a cartoon."

According to Randall C. Smith, staff research scientist, vehicle Analysis and Dynamics Lab, GM Research and Development Center (Warren, MI), he learned something about the importance of clay so far as designers are concerned, even though he is a computer scientist: it's sheer physicality. The clay model is there. An object in the design studio. Something that can be walked by from various angles. Something that is there in the morning and there at night. Something that can be assessed as an object. Smith talks about "soak time." Over a period of time, the designers, platform engineers, executives, and other interested parties who interact with the model are better able to perceive what are perhaps anomalous aspects that they would otherwise not see on a CAD screen... no matter how big the screen is. "We kept building bigger and bigger displays," Smith says. "People wanted to see them [the vehicles being developed] full sized." But he notes that there are some "fundamental perceptual issues." The thing is this: even though the rendering on a CAD screen, even a BIG screen, might be almost trompe ('oeil photo realistic, it is still in a comparatively contained space, one described by the dimensions of the screen. The clay model, on the other hand, is there, out in the environment, not in a separate plane of space, as the CAD digital model is. Even if a car is displayed on a typical screen 24/7, the same sort of soak can't occur with CAD... at least not until Smith and his colleagues are done with it.

There are two basic approaches to displaying the information. One is on the aforementioned screen, which is a huge rear-projection unit that, in the case of the GM Design Studio in Warren, is of a width that dwarfs what you are likely to find in a Cineplex. The other display is called the "CAVE," which stands for "Cave Automatic Virtual Environment." In the CAVE, the images are projected on three walls, thereby providing the viewer with a sense of being in something.

The hardware for driving the images consists of Silicon Graphics Onyx2 computers. In the lab, for example, the hardware is setup in so that there are what amount to four workstations driving the images for the CAVE. Because there is a need to rapidly redraw the images so that the viewer has the sense that she is actually looking at something even though she moves her head from point to point on the object (there is actually a buffer required to allow perceptually seamless transitions from one view to the next), the data-crunching capability of the hardware is essential.

To achieve the 3D effect, viewers are fitted with a set of glasses that have liquid-crystal lenses. Essentially, the glasses open and shut very rapidly; there is no massaging of the data, but the math models displayed are perceived at two slightly different views, thereby providing the sense of depth. With regard to the software, data from both Unigraphics CAx software and from Alias rendering software are loaded into VisualEyes, and transformed. One of the objectives of the development was to make the system useable by someone who doesn't have a Ph.D. in computing. (GM is passing VisualEyes technology to Engineering Animation, part of Unigraphics, so that deployment throughout GM's design organization will be facilitated, as the company is familiar with creating interfaces that can be used by the engineering community.)