Comparative study of target acquisition performance between an eye-slaved helmet display and unaided human vision

Abstract

This paper presents the results of a comparative study on the target acquisition performance of an eye-slaved helmet display system and unaided human vision. A teleoperated visual sensor system records the visual information that the human visual system can sense at a remote site, transmits the information to a remote operator, and reconstructs the information in real time on an eye-slaved helmet display system. The display system design is based on the following: the physiology of the eye shows that the density of cone cells responsible for scotopic vision reduces drastically with off-coresight angle, as well as the corresponding resolution. The human visual system can be adequately stimulated over a 60 degree field of view with only two video channels, a wide field of view for peripheral vision and a narrow field of view for foveal vision. The wide-field and narrow-field information can be displayed into separate eyes with the human brain doing the integration between the two pictures. This paper provides experimental proof of this hypothesis. The narrow field display can be made to track the human eyeball during target acquisition tasks, without unduly distracting natural vision. During target search tasks the human eye movement exhibits rapid saccadic movements after which it fixates on some feature on the picture. During and for a small time after the saccadic movement,the neural activity of the eye is inhibited. This inhabitation period can be used by a machine to track the eyeball and to slave the picture to the eyeball through a set of mirrors, without injecting conflicting information to the human visual system during the saccade. An eye-slaved helmet display system was built with a simulated monochrome target scenario. The final experiment evaluates the target acquisition performance of a number of subjects with both the eye-slaved helmet display system and direct vision on the same pre-recorded scenario. The experimental design used two different target contrast, four target ranges and four positions. Two control experiments were also done, one in which the same scenario was printed on a viewing board and a set of target acquisition experiments done with unaided vision, and the other where a joystick controlled sight was simulated with a terminal display system. The eye slaved display result is better than the terminal display result, and not far off from the direct vision test results. It was found that test subjects had to be free of strabismus and monoscopic fixation, for the eye-slaved helmet display system to work. We conclude that an eye-slaved display system design is the key technology for remote controlled sighting systems with similar capabilities to unaided human vision.