|
|
 |
|
|
|
| Visual Performance and Fatigue in See-Through Head-Mounted Displays |
|
|
Principal Investigators
Stephen R. Ellis, NASA Ames Research Center
Bernard D. Adelstein, NASA Ames Research Center
Abstract
Head-mounted see-through displays have been used to examine the cause of errors in human observers' depth judgments of computer generated virtual objects. Interactions between the virtual images of the virtual objects and the real surfaces over which they are optically superimposed have been studied. Changes in static convergence have been associated with errors in operators' judged depth of the virtual objects as indicated by adjusting a pointer. The consequences of monocular, biocular, or stereoscopic viewing on the accuracy of these depth judgments and on the subjective viewing discomfort during extended periods of time have been studied. The displays used in these studies promise to dramatically increase productivity in several manufacturing environments and to be unique stimulus presentation formats for scientific research. The amount of measured full system rendering delay for the presentation of head-stabilized, stereo-scopic virtual objects has, however, remained a significant problem for the use of these displays. Recent development under this project has reduced the full system delay from 65 msec to 30 msec and correspondingly trebled of average rendering update rate to 60hz. This performance is currently among the best dynamic response of a rendering system like this in the world and has been used to assess the latency requirements for systems that present virtual objects of variable complexity and which require variable precision of operator interaction.
|
Click Image to watch video
(Quicktime - 3.7MB)
|
|
Fiscal Year 1995 Task Progress
The three principle accomplishments of the head-mounted see-through display task in the last year have been
1) the completion of four experiments examining the cause of errors in depth judgments to virtual targets presented via head-mounted displays; 2) The completion of two experiments examin-ing the consequences of monocular, biocular, or stereoscopic viewing on the accuracy of depth judgments of virtual objects. Subjective viewing discomfort while using the displays was also studied; 3) The reduction of measured full system rendering delay for the presentation of head-stabilized, stereo-scopic virtual objects from 65 msec to 30 msec and corresponding trebling of average rendering update rate to 60hz.
These accomplishments have been reported in refereed proceedings papers and are under consideration for publication in refereed journals.
1) Four experiments have examined a displacement of the judged position of a nearby virtual object associated with the superposition of the object against physical surfaces within arms reach. The experiments measure the extent of the error in depth judgment and show the error is associated with changes in static convergence induced by the presence of the physical surface. The experiments distinguish between purely oculomotor explanations and those based on perceptual interpretation of occlusion. They suggest that shearing of optical contours may be a sufficient stimuli to induce convergence reflexes.
2) Since stereoscopic presentation of virtual objects generally involves rendering for both a left and right image, there is considerable practical gain from examining monocular head mounted displays that may be adequate for many tasks. The monocular systems only require one rendering and one set of viewing optics so they are inherently cheaper and lighter. Studies comparing the accuracy of depth judgments with monocular, biocular, and stereo viewing have shown that for static images with appropriately adjusted convergence planes, biocular displays can present virtual objects with high judgment accuracy comparable to stereo displays. Studies of viewing fatigue while visually tracing space stabilized virtual objects; however, show that biocular viewing conditions produce significant, unique viewing difficulties probably due to conflict between looming and disparity cues to ocular convergence.
3) Excessive end-to-end latency and insufficient update rate continue to be major limitations of virtual environment (VE) system performance. Improved hardware and software reconfigurations have reduced end-to-end latency and increased the update rate. These reconfigurations included: 1) multiple asynchronous UNIX processes communicating via shared mem-ory; 2) continuous streaming rather than polled tracker oper-ation; 3) multiple rather than single tracker instruments; and 4) higher bandwidth IEEE-488 parallel com-munica-tion between tracker and computer replacing RS-232 communication. Average latency of 65 msec and an update rate of 20 Hz for a standard 1000 polygon test VE, has now improved to 60 Hz (the maximum achievable with our graphics display hardware) with approximately 30 msec average latency. Because our equipment and architecture is based on widely available hardware (i.e., SGI computer, Polhemus Fastrak) and software (i.e., Sense8 WorldToolKit), our techniques and results are broadly applicable and easily transferable to other VE systems. Three dimensional tracing of virtual objects resembling smooth and segmented wire paths has been studied with system latencies varying from 48 msec to 500 msec. Operator performance degraded with increased latency, as expected, but showed an interactive effect of required precision and path complexity. Higher precision complex paths appear to require latencies less than 48 msec for optimal performance.
Results from accomplishments 1 and 2 have been provided to Boeing Computer Services and GM Technical Centers in response to requests. Code from accomplishment 3 has been made available through appropriate NASA distribution systems when requested. Current investigations are examining role of motion parallax in aiding depth judgments with existing dynamic response and those enhanced by several alternative predictive tracking systems. Improvements in the full system dynamic response of computer graphic based simulations presenting virtual objects on head-mounted displays (HMDs) are being provided to improve dynamic performance of HMDs to be used for NeuroLab experiments led by Dr. Charles M. Oman of MIT.
Recent Publications
- Ellis, S. R. and Menges, B.M.(1996) Effects of age on judged distance to virtual objects in the near visual field. to be presented Proceedings of the Human Factors and Ergonomics Society, 40th Annual Meeting. Philadelphia, PA.
- Noor, A. K. and Ellis, S. R. (1996) Engineering in a virtual environment. Aerospace America July, 32-37.
- Jacoby, R. H., Adelstein, B. D., and Ellis, S. R. (1996) Improved temporal response in virtual environments through system hardware and software reorganization. Proceedings of the SPIE 2653, Stereoscopic displays and virtual reality systems III February, 1996.
>- Ellis, S. R. (1995) Virtual experience, Invited lecture; 1995 Meeting of the Optical Society of America. Sept. 13, 1995 Portland, Oregon.
- Ellis, S. R., (1995) The design of pictorial instruments. In Human Computer Interaction and Virtual Environments, NASA CP 3320, NASA Langley Research Center, pp. 13-24.
- Ellis, S. R. & Menges, B. M. Judged distance to virtual objects in the near visual field Presence (accepted for publication)
- Ellis, S. R. & Menges, B. M.(1995) Judged distance to virtual objects in the near visual field. Proceedings of the Human Factors and Ergonomics Society, 39th Annual Meeting. San Diego, CA, 1400-1404.
- Ellis, S. R. (1995) Presence of mind: a reaction to Sheridan's musings on telepresence. Presence 5, 2, 247-259.
- Ellis, S. R. (1995) Human engineering in virtual environments. Virtual Reality World, '95 IDG Conferences & Fraunhofer Institutes IPA IAO, Stuttgart, Germany, pp. 295-302.
- Adelstein, B. D.., Johnston, E. R., & Ellis, S. R. (1995) Spatial sensor lag in virtual environment systems. Presence (in the press)
- Ellis, S. R. , Bucher, U. J. & Menges, B.M. The Relationship of Binocular Convergence & Errors In Judged Distance to Virtual Objects. Proceedings of the International Federation of Automatic Control June 27-29, 1995 Boston, Mass.
- Ellis, S. R. (1995) Virtual environments and envi-ronmental instruments. In Simulated and virtual realities, K. Carr and R. England, eds. Taylor and Francis, London. pp. 11-51
|
|
|
