A group of scientists and educators from NASA Goddard Space Flight Center, Purdue University, and the Howard B. Owens Science Center are working together to bring the classic Powers of Ten concept of moving between worlds of different magnitudes into personal computers in an interactive medium. The project, known as the Interactive Universe, is targeting the regions from the scale of the Earth up to the universe. The goal of this project is to help students, teachers, and the public conceptualize and enjoy learning about the structure, size, and contents of the universe.
The advent of the World Wide Web (WWW) and the Virtual Reality Modeling Language (VRML) makes this possible with tools to create a representative Universe and deliver it to the desktop. VRML is an emerging technology for 3D modeling and interaction over the Web. VRML is a descriptive language used to model 3D objects such as planets, stars, galaxies, people, buildings, and molecules and to allow the user to control the viewing perspective of the object. A VRML object may be defined using a combination of basic geometric primitives (e.g., cube, sphere, cone, or cylinder), or a complex arrangement of polygons. The software required to view VRML models is available for free at a number of locations on the Web such as the VRML Repository. The VRML viewer may be a plugin to an Internet browser (e.g., Cosmo Player) or a standalone application (e.g., VRWeb) with navigation aids to moving around or zooming in/out of the scene.
The Interactive Universe Project
The Interactive Universe Project is divided into six domains, each of which is composed of single or multiple VRML models with associated documentation and related information.
|Solar System||39.48 AU (Sun/Pluto distance)|
|Nearby Stars||25 pc|
|Milky Way Galaxy||30,000 pc|
|Local Galaxies||3,000 kpc|
|Universe||50,000 kpc (*)|
|(*) Although the size of the known Universe (i.e. distance to the furthest galaxy) is approx. 10 billion light years, this preliminary model is considerably smaller.|
The VRML models allow an observer to fly through a given area and see the scene from any position. Each group of models is described in the following sections:
The Earth-Moon model depicts the Earth and Moon to scale with the relative distance between them also to this scale. Figure 1 shows a screenshot of the VRML model and Figure 2 is an actual photo taken from the Lunar 3 Orbiter (available from NSSDC's Image Catalog). This model is essentially two spheres texture-mapped with appropriate Earth and Moon pictures. The latest version of VRML, VRML 2.0, allows motion to be included such that the Moon can orbit the Earth. The actual orbital period of one month could be simulated, but, of course, it would not be visually noticeable. The Moon's speed can be accelerated so the motion is apparent, but at present this has not been incorporated.
|Figure 1. VRML model||Figure 2. Actual photo|
The solar system, like the Earth-Moon model, is represented using colored and textured planetary spheres of the proper scale, size, and relative distance from each other. However, the actual solar system is hardly viewable given the minuscule size of the planets in comparison to the vast distance among them. To overcome this problem, a CGI script has been created with options to magnify the sizes of the planets and Sun while keeping their separations fixed. In this manner, an observer can first see the solar system as it really is and then slowly increase the magnification to see the planets better. There is also an option to create Sun-sized cube outlines around each of the planets to show clearly the position of the planet, which would normally appear only as a speck on the screen. Some models include solid-colored spheres while other more realistic models include texture maps, which are actual photos of the planets wrapped around a spherical object. Figure 3 is an example of this with Venus as a textured planet and the Sun in the background.
Figure 3. View of Venus in front of the Sun
The nearby stars model represents the 3,589 stars within 25 pc of the Sun. The sky coordinates (right ascension and declination) of the stars, distances, and spectral types were taken from the Gliese's Third Catalogue of Nearby Stars. The coordinates and distances were used to calculate cartesian (x, y, z) coordinates in a three-dimensional grid with the Sun at the origin (0, 0, 0). A "realistic" color was assigned to each star based on its spectral type (e.g., our Sun is a G2 class star represented by a greenish-blue color). The stars are sorted by the color into 24 groups where each group represents a different class of stars. The result is a spherical blob of colored points. (See Figure 4.)
Figure 4. Nearby stars
The Milky Way Galaxy has over 100 billion stars. The VRML model for this would exceed the memory capabilities of most desktop computers, so it was decided to use Monte Carlo sampling of realistic star density distributions based on the work of Wainscoat et al. (1992) and Dwek et al. (1995). VRML models are available with 40,000 and 4,000 stars with a realistic distribution among 15 spectral types. The two models offer a trade-off between more realism and a smaller file size. The more detailed model is shown in Figure 5. It was also necessary to enhance the spiral arm star density by a factor of 20 to make the arms visible.
Figure 5. Top view of Milky Way Galaxy
The local galaxies model was created using Tully's 1988 Nearby Galaxies Catalogue along with the Third Reference Catalogue of Bright Galaxies (RC3). The actual position, size, orientation, and type of galaxies are represented in this VRML model for the local galaxies with the Milky Way at the origin. Fifty-five galaxies at distances of up to 3,000 parsecs are included in this model, divided into four types: spiral, elliptical, lenticular, and irregular. The Andromeda Galaxy M31 (NGC 224) with its two companions (dwarf elliptical galaxies M32 and M110) are shown in Figure 6.
Figure 6. Andromeda galaxy with its two companion galaxies
As with the nearby star and Milky Way models, the universe model is a collection of star-like points of light, but here each point represents a galaxy. (See Figure 7.) The model uses the same Galaxy catalog as for the local galaxies and includes over 2,000 objects within a distance of 50,000 kpc from our galaxy. Larger catalogs (e.g., the CfA and Las Campanas redshift surveys) will be used in the future with more galaxies at greater distances included.
Figure 7. At the heart of the "universe" -- the Milky Way
Traveling through a virtual universe of such size can be difficult since much of the scene looks alike and there are not well defined reference points. The result of such an exploration could be a feeling of disorientation; that is, these models give new meaning to the term "getting lost in cyberspace." Therefore, to navigate such spaces, reference points are needed.
One approach to help students find their way around the models is to provide viewpoints with predefined views into the scene from a specific location and perspective. The observer can return to any viewpoint from a menu on the VRML viewer or cycle through each viewpoint for a quick tour of the model. The viewpoints help to navigate around the model and get back to an area of interest, but it is best to venture out from the viewpoints and explore the objects from various positions.
Early Problems with VRML
Although VRML in theory is portable across all computers the implementations vary greatly. One model works well with one browser/platform combination but poorly with another. Texture maps and transparency are handled very differently from a SGI to a PC implementation (i.e., the SGI workstation has specialized graphics hardware that is not present on PCs). Some models such as a large Milky Way Galaxy with over 50,000 stars tend to be either too slow to navigate or take up too much memory and crash the VRML viewer upon display. Also, having too many viewpoints (e.g., >50) in a model can crash some viewers.
The Interactive Universe Project was conceptualized shortly after the first VRML viewer was publicly released and has evolved as VRML itself has evolved and matured. The project is far from completion with the creation of on-line documentation and a lesson plan/curriculum for K-12 students based on the material still to come. Further plans include evaluating students' experiences with the current models and incorporating their feedback into revised models. VRML has made visualizing the planets and the universe practical for low-end PCs that are commonly available in many schools and homes. There are many shortcomings in VRML that need to be addressed, especially cross platform compatibility issues, but the future of VRML holds great promise as the language and software are improved.
Interactive Universe Website
NSSDC Image Catalog Lunar Orbiter Photo #L04-M123
Erin D. Gardner, email@example.com, (301) 286-0163
Raytheon STX, Code 633, NASA Goddard Space Flight Center
Greenbelt, MD 20771, U.S.A.