|
|
 |
|
|
|
| High Fidelity Virtual Environments |
|
|
Principal Investigator
Stephen R. Ellis
Abstract
Virtual environments are personal simulators. They are interactive, head-referenced, computer displays that create an illusion that their users are displaced to another location. This illusion is created through the operation of three types of equipment: 1) Sensors, such as head position sensors, to detect operators' body movements , 2) Effectors, such as a stereoscopic displays, to stimulate the operators' senses and 3) Special purpose hardware to interlink the sensors and effectors to produce sensory experiences resembling those encountered by inhabitants immersed in a physical environment. In a virtual environment this linkage is accomplished by a simulation computer. In a head-mounted teleoperator display the linkage is accomplished by the robot manipulators, vehicles, control systems, sensors and cameras at a remote work site. Both virtual environments and head-mounted teleoperator displays have applications in mechanical design, data visualization, robotics and aids for mechanical assembly.
Current virtual environment systems commonly use electromagnetic position trackers to sense operator head and hand positions to generate virtual environment simulations. The most common of these sensors suffer from uncorrected distortion in measurements of operator position and when interfaced to simulation computers produce objectionable time lags. Though these defects acknowledged problems, their impact on the objective and subjective aspects of operator behavior within immersing virtual environments has not been studied and they have not been minimized.
The objective of the current work has been to measure and correct the position distortion of common position sensors (Figure 1) and to minimize simulation system visual lag during virtual environment rendering (Figure 2). These improvements are then tested with a manual tracing task in which operators attempt to move a virtual ring over a virtual path (Figure 3 left & center) without making contact with it. Path complexity and ring diameter are changed (Figure 3 right) to study the precision with which operators can accomplish the task as a function of display conditions.
The precision and accuracy with which users may interact with virtual objects in virtual environments has been improved through correction of spatial distortion in common position sensors users to produced virtual environment simulations and by improvements in the latency and update rates with which these environments may be created. Position and orientation of a FasTrak position sensor were measured and corrected comparing linear and nonlinear interpolation schemes adapted from the computational geometry used by Computation Fluid Dyanmics. These algorithms have been ported to run on popular computer graphics workstations.
Human performance studies are being conducted to compare objective and subjective performance within virtual environments using optimized simulation systems. Operators' three dimensional tracking error, adapted Cooper-Harper controllability scales, subjective rating of visual stability, feelings of nausea and the accuracy of depth rendering due to motion parallax are being used to evaluate the benefits of optimized immersing environments and head-mounted displays of virtual objects.
|
|
|
|
|
