 |
 |
 |
|||
 |
 |
 |
 |
|
|
 |
|
||||
 |
|
||||
Images
& Movies
Posters
Talks
Publications
Tools Research
3D Force Microscope
Mixing Model / Experiment
NanoManipulator
Advanced Technology
Collaboration
Biomedical Research
Cystic Fibrosis
Fibrin and Blood Clotting
Gene Therapy and Viruses
DNA
Cell Division
Bacterial Motility
Molecular Motors
Nanoscale Science Research Group Home
Advanced Technologies Research
Because
our mission involves bringing the best of new computer technology to bear
on biological problems, we have always invested between 10% and 20% of our
effort in technological exploration of a high-risk, high-potential payout
sort. This we call our advanced technologies research. It has paid off quite
well.
A significant fraction of these explorations have come to naught. For the ones that turned out well, our usual pattern has been to spin that research off as a separate project, after it has been shown promising enough to attract its own research funding support. The big exception, of course, is the nanoManipulator project, which, after it proved out, we have adopted as a main thrust of the Resource, phasing out the molecular structure work. Here we briefly list the wildcatting research that has spawned separate projects (or sub-disciplines):
- Volume Visualization. The Resource developed the three major techniques used to directly display volume data (interlaced isocontours, volume ray-casting, and volume splatting) and hosted the first Conference on Volume Visualization.
- Virtual Environments. On the way to displaying a room-filling molecule, from the early 80's we have been early adopters and innovators in virtual environment systems. This re-search has led to a group of local researchers who have a history of being world leaders in the technology and application of VE systems. [Robinett1992; Mine1997]
- Head-Mounted and other Display Devices. When early HMDs were of too low resolution to display whole molecules, we designed and built high-resolution models; to form a shared environment for two seated users we built the projector-based PIT system; to provide the ideal workbench for nanoscale manipulation, we built a rear-screen projection system. [Arthur1998; Grant1998]
- Tracking Technology. When room-filling trackers were unavailable, we began to explore wide-area optical tracking technology that has since found its own funding and is being commercialized.
- Algorithms for Real-Time Rendering. We developed a patented
algorithm for rapid dis-play of spheres; implemented in microcode on
three generations of world's fastest graphics computers. In order to
break the barrier of 150K triangles per frame on high-end hardware,
we have investigated a wide range of techniques (culling, simplification,
image substitution and prediction) that increase the realism of rendered
scenes while decreasing their rendering time. [Aliaga1996; Cohen1996;
Erikson1996; Luebke1996; Luebke1997]
Advanced technology research is still going on. The Resource currently supports a modest effort in the basic science of graphical display in which one is "present in" the scene, rather than "looking at" the scene. We have noticed informally many times in the past that users are able to ignore even egregious errors in the display system when they are concentrating on a problem; they are some-how on task. Indeed, one goal of the nanoManipulator interface is to enable the scientist to feel present at the nanometer scale, rather than feel as if they are interacting with it from a distance (compare the difference between exploring from Earth using a link to the Mars Lander with be-ing present on the surface of Mars.)
Several groups have begun to study the effects on perceived presence of different characteristics of VE systems. [Barfield1993; Hendrix1996; Büscher1998; Durlach1998; Slater1998; Fen-cott1999; Usoh1999]. We are asking ourselves how VE presence and being on task are related, and how best to build systems that encourage this. What factors affect presence? What are the system characteristics that will let the scientist attend to his science and not to the display or to the tools?
Our current experiments capitalize on our having serendipitously built an exceedingly compelling virtual experience, which creates strong fear-of-falling reactions. We now instrument our subjects to measure heart rate and palm sweat, yielding an objective, robust, and fine-grained measure of presence. We are backing the quality factors of our environment off, one by one, to see which factors are essential for getting strong presence reactions. So far, we have studied the effects of static haptic cues and of changing frame rate.
We have completed two studies. The first study tested the ability of the physiological measures noted in the previous section to discriminate between two conditions, with and without the static haptic models. The results showed that the measures discriminated between the two groups at a significant level and that the measures correlated strongly with presence as measured by observed behavior and self-report in post-experiment questionnaires. This study validated our use of physiological measures as a correlate of presence. Further, it showed that virtual environments that include crude physical models of real objects in the environment that the subjects can actually touch and run into (static haptics) induce a higher sense of presence than those that do not include such models.
The second study tested the hypothesis that user sense of presence is directly related to the rate at which the visual data presented to the user is updated (frame update rate); the higher the update rate, the higher the sense of presence. Each user performed the experimental task at four (4) frame rates. As before, we are comparing physiological, behavioral, and self-reported presence measures. This is the first in a series of studies in which we will examine the technical factors of our environment, as noted above.
For more information see the Effective
Virtual Environments site.