Interest in optimized human-computer interfaces has resulted in the development of a number of interesting devices that allow the computer and human operator to interact through a common drawing surface. These devices include the lightpen, lightgun (Goodwin, 1975), and a variety of touch-sensitive display overlay devices. Although touch devices were being investigated as early as 1965 (Orr and Hopkin, circa 1966), behavioral and performance data are scarce in relation to other sources of human-machine interface data. Availability of these devices has increased in the last 10 years and it is now possible to retrofit such devices to a wide variety of video display terminals at a reasonable cost. With the possibility of increased use looming on the horizon, it would be quite useful to examine the ergonomics of such devices and the behavioral adaptation or maladaptation that occurs for each user. Performance data available at this point from previous studies suggests that some positive increments in performance can be expected for graphic-based tasks while no serious decrements should be expected for discrete data entry tasks (Beringer, 1980; Stammers and Bird, 1980). The performance gains expected from this format of interaction are not to be won without some sacrifice elsewhere, however. Positioning of the display surface for optimum viewing may cause serious operator fatigue problems after extended use of the device if the device is to be used with relatively high frequency. The relationship of device positioning, device sensing resolution, and task type are being examined as they contribute to the comission of errors and the onset of fatigue. Experimentation was planned to examine how positioning of the device, or what can truly be called a “control/display unit”, affected the performance of visual discrimination tasks and manual designation tasks. Initial investigations used a single task to examine these questions by requiring the operator/subject to visually detect and manually designate the location of a break in one of 54 circles presented on a color c.r.t. display (essentially a Landholt C target). Responses were accepted by an infrared touch panel mounted on the display face. The c.r.t. was placed at four declinations during the blocks of trials; 90, 67, 45, and 35 degrees to the line of sight. Although a very strong learning effect was observed over the first 8 blocks of 25 trials each, performance leveled off, on the average, beginning with the ninth block of trials. No reliable effects of screen declination were found in the examination of response times or number of errors. Responses did tend to be located slightly lower than the target, however, for the greater declinations of the display surface. Subjective reports of physical difficulty of responding and fatigue did vary regularly with declination of the display. The relatively high location of the device resulted in shoulder and arm fatigue when the display was at 90 degrees and wrist fatigue when the display was at 35 degrees. Subsequent phases of the investigation will allow subjects to adjust parameters of height and declination (Brown and Schaum, 1980) and will use hand skin temperature and quantified postural information to assess the degree of fatigue incurred during device operation.
EEG-Based BCI Emotion Recognition: A Survey