#51 from R&D Innovator Volume 2, Number 8          August 1993

Virtual Innovation
by Brad Burnett

Mr. Burnett is director of R&D at VRontier Worlds, Inc., Stoughton, Wisconsin.  He contributed a chapter to Adventures  in Virtual Reality, by Thomas Hayward (Que, Carmel, IN, 1993).

My strange trip through the world of the almost-real began about a year ago, when I was running a home electronic-design company and getting an associate degree in electromechanical design at a vocational college.  The catalyst that changed my life was a 10-minute segment of a TV documentary about the evolution of computers.

That particular portion showed a man wearing a large helmet.  Right in front of his eyes, two video screens showed the inside of a building that had not yet been built—a building that did not exist.  A small tracking device on this virtual reality (VR) helmet caused the computer-generated image to change in accordance with the position of the viewer's head, just as a scene changes in the real world when you move your head. 

Simply, the video screens and optics in this head-mounted display allow the wearer to "enter" a computer-generated world.  Unlike the Viewmaster stereo slide viewer I remember from childhood, there are no edges around the image; you don’t see the "world" through "holes" in the device.  Instead, VR helmets give you the sensation of being inside the world and looking out.  The device occupies most of your field of view, and gloves, joy sticks, and head-tracking devices allow you to interact with the "virtual space" in a natural way.

A Benign Obsession

I was instantly obsessed with the desire to invent a better display than the 10-pound clunker shown in the program, so I requested information from sources listed in the credits.  Since nobody was willing to tell me much, and it appeared that the people I contacted at various university labs and VR companies didn't have a great deal of optical expertise anyway, I realized I would have to design a head-mount system from scratch.  At this point, I was only thinking about making a system for myself—not a commercial product. 

My first headmount was built around two liquid-crystal video display screens I extracted from miniature televisions.  I got the electronics to work immediately—but that's the easy part of a headmount.  The virtual reality industry has been struggling with optics for years, and still focuses on a headmount lens system developed for NASA in 1985 to view photographic slides. 

That system was black-and-white, a real dinosaur in video terms.  A second problem was the weight of the optical package—about 1.2 pounds.  Because the weight in the front of the helmet must be balanced by a counterweight in the back, each ounce saved in optics and electronics removes two ounces from the user's head.

But the fatal flaw of this early design arose from simple geometry:  When the smallest available video screens were mounted side by side, their centers were about 3.2 inches apart, far wider than the span between an adult's eyes, 2.5 inches or less.  Because human eyes cannot diverge, most existing VR systems use software to place the images directly in front of the eyes.

Nevertheless, I needed a starting place, so I ordered the expensive optics used in the NASA viewer.  Although they worked beautifully on the photographic slides they were designed for, they left a lot to be desired when used to focus video screens.  Progress required a lighter optical system that could place the screens directly in front of the eyes, remain in focus for different users, and have a large enough field of view so it would seem to "wrap around" the user.

Learning on the Job

I'm an electromechanical designer with no education in optics.  But I do possess two skills that help me challenge the unknown.  I am willing to put out the energy to educate myself in areas that interest me.  More important, I seem to have an intuitive ability to solve difficult problems without having to wade through all the details and explore all the dead-ends.

I started by convincing someone at an optical supply company to put the lenses of the NASA headmount under an instrument, so we could measure their optical properties and begin to understand what was going on.  I also went to the library to learn the jargon and principles of optics.

If I had the misfortune of a formal education in optics, I probably would have missed my solution.  But it was precisely my ignorance that allowed me to transcend the barriers of the conventional.

I stumbled onto some flat, vinyl Fresnel lenses.  These lenses have multiple concentric rings and are used in overhead projectors and the wide-angle lenses found in the rear windows of campers.  I thought a compound optical system (one lens mounted in front of another) would allow lighter weight and higher power, but the conventional wisdom held that compound Fresnel lenses would suffer from moiré patterns, the wavy lines that form, for example, when one window screen is superimposed on another. 

Nevertheless, my first prototype with compound Fresnel lenses did things conventional lenses never could: it wrapped the scene around the user's head and adapted easily to different users.  And because I was using liquid crystal displays rather than video screens, I considerably reduced weight and electric current draw.  Most important, the assembly is small enough to sit directly in front of the user's eyes.

We Met on a Bulletin Board…

After some media coverage of my development and a lot of communication with others around the world, I met, through a computer bulletin board, two gentlemen who had a small virtual reality software company called VRontier Worlds.  I was surprised to find anyone working with virtual reality in my home state, Wisconsin.  We became equal partners in the firm, which sells one of the few VR software packages able to run on Intel 486-based personal computers. 

The headmount project, which has been consuming 17 hours a day of my time, continues its rapid advance.  My streamlined optical system has fewer parts, allowing faster assembly and a more comely appearance than the Frankenstein-bolt-through-the-head look found in competing headmounts.  We use custom Fresnel optics  made of a highly transparent plastic, cut to extreme accuracy, to get a satisfactory view.

I also began a patent search to be sure I wasn't infringing anyone else's discovery.   And I began to optimize the inevitable tradeoffs of optical design.

To demonstrate the headmount, we run our own VR program showing one airfield, two buildings and a pair of hot-air balloons.  If you are using the joystick, when you pull back and press the button, you're climbing.  Push it forward, and you move forward but altitude remains constant.  Tilt the stick left, and your view sweeps left.  If you want to look inside the building, just go "through" the wall, and check out the interior from any angle you like.  If you're using the head-tracker, you get similar results just by moving your head.

Today, less than a year since I started work on the headmount, my device has already been tested by NASA with good results .  McDonnell Douglas has also used it in a mockup of a Mars probe, which would use a pair of cameras to transmit a realistic view of Mars to Earth (and into the headmount).

In one month of shipping these headmounts, we have sold more than 10—not bad considering they retail for nearly $10,000.  Our customers include virtual reality hardware and software companies and university psychology researchers.  We are collaborating with an architect who will use the system to show virtual buildings to clients.   I can envision surgeons training by taking a trip through a virtual body or technicians practicing hazardous-waste cleanups in the safety of a classroom.

Now, with some money coming in, I'm working on version two of the headmount, which will have greater visual acuity and a wider field of view.  I think my unconventional understanding of Fresnel lenses will allow me to combine these hitherto incompatible goals.  We may use our own VR system to design the next headmount—a sort of computerized birthing process.