Although the hardware and software for virtual environment systems have rapidly progressed during the past few years, many theoretical and practical human factor issues on the systems remain to be explored. Among the issues, the enhancement of human performance in virtual environment systems is especially important in some application fields such as architectural navigation, vehicle simulation, and teleoperation.
In this study, a series of four experiments were conducted to investigate human performance variations of the three-dimensional pointing task in a virtual environment. The first two experiments deal with the effects of control factors, while the others deal with the effects of display factors.
The first experiment was prepared for the following experiments, in which the display/control (D/C) gains of 6 degree-of-freedom controllers were optimized. The second experiment was designed to validate Fitts`` law, which has successfully described the one- or two-dimensional control tasks, for the three-dimensional pointing task.
The results of the experiment showed that Fitts`` law did not describe the three-dimensional pointing task well compared with the two-dimensional pointing task. The poor fitness was discussed in relation to the characteristics of three-dimensional pointing tasks. The high degree of freedom in control, the egocentric viewpoint change, the poor display conditions, and the clutching problem of the three-dimensional pointing task decreased human performance and increased its variation. For the clutching problem of isotonic controllers, a modified model of Fitts`` law was proposed. It improved the coefficient of determinant (R2) of the data.
In the third experiment, three display factors (frame rates, screen resolution, and pictorial contents) were optimized to improve human performance. The results showed that the fixed frame rate at 18Hz was superior to the higher unlimited frame rates. Also, high screen resolution and simple visual cu...