Search

Chunking is defined as the process of organizing information into multi-item clusters appropriate to solving a particular task. It was hypothesized that dynamic information displays could be formatted to group related items. This formatting would facilitate chunking and, in turn, serve to reduce task difficulty. Two novel display grouping formats were evaluated. In Tight Spatial Proximity, iconic symbols belonging to the same class were rearranged so that they were located proximally on the display. In Chunks-In-Sequence (CIS), symbol classes were presented sequentially, but all members within a symbol class were presented simultaneously. Ten individuals participated in Experiment 1. During a trial, 2 to 25 symbols (e.g., cannon, radar, etc.) were presented briefly on a computer display. Displayed symbols were sampled from a larger symbol set (Le., 64 total symbols) representing eight unique symbol classes (e.g., military, highway signs, etc.). Participants then viewed a prompt and responded "yes" or "no" to indicate whether the probe symbol was a member of a class that had been presented during that trial. Results indicated that grouping formats helped participants chunk symbols into appropriate classes and solve the classification task with reduced error rates and subjective workload. Experiment 2 was similar to Experiment 1, with the exception that participants were tasked to remember each symbol presented during a trial rather than just the symbol class names. This task required forming chunks, storing them in memory, and then parsing the chunks into component items. Results indicated that display grouping formats reduced subjective workload, but not error rates. Considering both experiments, it is concluded that spatially or temporally grouped symbols can be chunked more easily than information displayed in a non-grouped format. However, grouping formats will not help operators parse these information chunks given detailed component recall demands. Ph. D.

This dissertation examines those attributes of full-color display systems (particularly color matrix displays) which degrade image quality. Based on this analysis, it is suggested that a comprehensive metric should measure image quality in terms of transmitted signal and noise modulation, both achromatic and chromatic. Moreover, it is suggested that these signal and noise measurements be weighted in terms of human spatial-chromatic visual characteristics. A review of extant image quality metrics reveals several limitations of these metrics which make them unsuitable for the evaluation of color matrix displays. These limitations include the inability to account for chromatic modulation transfer and chromatic noise as well as the general inability to account for spatial and grey-scale sampling. This work describes a new methodology for assessing image quality that can be applied to full-color as well as monochromatic, and sampled as well as continuous, display systems. Unlike most display quality metrics, the proposed methodology is not based on the tools of linear systems analysis. Rather, it is based on more veridical models of the human visual system (HVS), including multi-channel models of spatial vision, the zone theory of color vision, physiological models of retinal processes, and models of the optics of the eye. A display evaluation system consisting of the HVS model used in conjunction with a display simulator is described. The HVS model employs nine image processing stages to account for nonlinear retinal processes, opponent color encoding, and multiple spatial frequency channels. A detailed procedure for using the HVS model to evaluate display systems is provided. The validity of the HVS model was tested by conducting contrast detection, discrimination, and magnitude estimation experiments on the model. The results of these experiments correspond closely with published human performance data The utility of the display evaluation system was assessed by making image quality predictions for the display systems used in three image quality studies. Image quality predictions using the proposed system correlate strongly with ratings of image quality provided by human subjects. Results of these validation studies indicate that the proposed method of display evaluation is viable and warrants further development. Ph. D.

An experiment was performed to evaluate the relations between active range of motion (ROM) and upper limb anthropometric dimensions. Eight anthropometric dimensions, forearm length, distal and proximal forearm circumferences, wrist breadth, wrist thickness, wrist circumference, hand breadth, and hand length in combination with gender, wrist position, and direction of motion or exertion were evaluated to determine their effects on instant center of rotation (ICOR) and the magnitude of force exertion. The knowledge gained from analysis of the study data will be the first step in the formulation of a biomechanical model of wrist flexion and extension. Such a model would predict forces and torques at specific wrist postures and be employed to reduce cumulative trauma disorders of the wrist. Sixty right-hand dominant subjects (30 male, 30 female) between 20 and 30 years of age all reporting no prior wrist injury and good to excellent overall physical condition, were employed in this study. The upper limb anthropometric dimensions and ROM were measured and recorded for each subject. The anthropometric dimensions were compared to tabulated data. The measured active ROM values were compared with values in the literature. Correlation coefficients between pairs of anthropometric variables (by gender) were calculated. The mean active ROM measures, 164.0 deg for females and 151.8 deg for males, were significantly different (Z = 2.193, p = 0.014). The relationships between the anthropometric variables and active ROM were analyzed by three methods: correlation between ROM and each anthropometric dimension, prediction (regression) equations, and analysis of variance (ANOVA). No correlation coefficient between ROM and any anthropometric dimension was greater than 0.7. No prediction equation, based upon linear and quadratic combinations of anthropometric dimensions variables, was above the threshold of acceptability (R² ≥ 0.5). The results of the ANOVA showed a significant effect for gender. The ICOR had been hypothesized to be either in the head or neck of the capitate. The Method of Reuleaux was employed to locate the leOR points for flexion and extension (over the ROM) of the wrist with three load conditions, i.e., no-load, palmar resistance, and dorsal resistance. Analysis of the data, using ANOYA, showed that wrist position was the only significant variable. Thus, in future wrist models, the assumption cannot be made that the wrist is a pin-centered joint for flexion and extension. The static maximal voluntary contractile forces that can be generated by recruiting only the six wrist-dedicated muscles in various wrist positions were measured. There was a significant gender difference for the mean flexion force (Z = 4.00, p = 0.0001) and for the mean extension force (Z = 4.58, p = 0.0001). Females averaged 76.3 percent of the mean male flexion force and 72.4 percent for extension. The force data, categorized by gender, were then analyzed using three methods: correlation of variable pairs, regression equations, and ANOVA. None of the eight anthropometric dimensions and ROM was correlated with flexion or with extension force at an acceptable level. The prediction equations, linear and quadratic combinations of all possible subsets of anthropometric dimension values, ROM, and wrist position did not meet the minimum acceptable level of R² ≥ 0.5. The ANOVA procedure showed gender, wrist position, direction of force exertion, and the wrist position interaction with direction to have significant effects upon maximal force exertion. Ph. D.

The measurement of software usability has become an important issue in recent years. Metrics of usability include time, errors, questionnaires, ratings, and results of verbal protocols. Concurrent verbal protocol, a method in which the user "thinks aloud" while completing given tasks, has been heavily employed by software usability researchers who want to know the reason a user is having difficulties. Possible problems associated with using concurrent verbal protocol are (1) that verbalization may interfere with the processing required to complete the task, and (2) that subjects may not be able to monitor and express the information of interest to the researcher. A relatively new approach which may avoid these problems is heavily cued retrospective verbal protocol in which the user is presented subsequently with a representation (a video tape, for example) which helps him recall his thoughts during the task without interfering with task completion. This research compared the performance of subjects while completing tasks using both methods of verbal protocol. The verbal data collected by the two protocol techniques was compared to assess any information differences due to the methods of collection. No performance differences were found between the two protocol methods. Reasons for this lack of degradation due to concurrent verbalization are discussed. The kinds of information gathered were quite different for the two methods, with concurrent protocol subjects giving procedural information and retrospective protocol subjects giving explanations and design statements. Implications for usability testing are discussed. The two methods of protocol were employed in a comparison of two different size monitors, a 30.48 cm diagonal and a 53.34 cm diagonal. The subjects' performance, as measured by steps to completion, task completion time, and errors committed, was compared across the two monitors. Subjects were required to complete 12 tasks which varied in the difficulty of the windowing required. Subjective data were also collected in the form of task difficulty ratings, as well as a global measure of user satisfaction. These performance measures and subjective measures were compared across protocol methods as well as monitors. Performance data, as well as subjective data, indicate that on tasks that do not require extensive windowing, there are no difference between the two monitor sizes. As windowing difficulty increases, however, the large monitor's advantages become apparent. Tasks with a high level of windowing difficulty are judged to be easier and require fewer steps on the large monitor than on the small monitor. Ph. D.

A television image is formed by a series of parallel luminous lines called a raster. The visual prominence of the raster structure interferes with the extraction of information from the image. The raster may be suppressed experimentally by a deflection process called spot wobble. Experiments were conducted to determine the effects of raster structure suppression on visual sine wave modulation thresholds and dynamic target acquisition performance using normal and noise degraded imagery. Results indicate raster structure suppression and improvements in Sine wave threshold sensitivity are correlated and that a suppressed raster significantly improves target acquisition performance for noise-free conditions. Performance correlations with the modulation transfer function area (MTFA) image quality metric were not as good as were the correlations between observer task performance measures and areas under the threshold functions. A rationale for improving the efficacy of the MTFA image quality metric was postulated. Ph. D.

A secondary task measure was used to evaluate the effects of various automobile handling characteristics on driver workload. The study was performed using the VPI & SU driving simulator, a variable speed random number generator and display, plus other control and recording equipment. Workload scores for 82 subjects were determined under various combinations of wind and curvature disturbances, vehicle steering gain, and vehicle response time. Driving performance was analyzed to measure intrusiveness of the secondary task and to confirm workload differences. The results showed definite differences across the levels of combined wind and curvature. These supported the assumption that the secondary task measure would be capable of measuring differences in driver workload. No significant differences in workload scores were obtained for changes in steering gain or vehicle response time. The secondary task did intrude on the primary task, indicating that its use may cause experimental complications. The method and equipment appear to be useable in either simulator or on-the-road vehicles. Master of Science

To determine whether incomplete reward reduction and reward increment would influence the "frustration effect" (FE) in double alleyway, rats were given 153 trials with 50% partial reward in Goalbox 1 with 4 or 12 pellets in Goalbox 1 and 4 or 12 pellets in Goalbox 2. On trials 106-153, the postshift phase, half of the rats in each group were shifted to a reward magnitude opposite to that previously received in Goalbox 1. During preshift, Alley 1 start and running speeds were inversely related to Goalbox 2 reward magnitude. The size of the FE was directly related to Goalbox 1 and Goalbox 2 reward magnitude. During postshift, Alley 1 running were inversely related to preshift Goalbox 1 reward magnitude. The size of the FE was directly related to Goalbox 1 and Goalbox 2 reward magnitude. The results were discussed in terms of Amsel's theory of frustration. Master of Science

Little research has been conducted dealing with human performance in symbol search of cartographic images on matrix-addressable displays. Alphanumeric studies have researched failures and polarity on such displays, but not using cartographic images. Two studies were conducted which required the subject to search for symbols on cartographic images under various conditions of failure, polarity, and background clutter. Response time and accuracy were measured. Generally, high clutter impeded performance more than low clutter, and negative polarity produced slightly better performance than positive. Horizontal line failures affected search ability more than vertical line or cell failures, particularly when the failures were in "on" mode rather than "off" mode. Further, it was found that while previous alphanumeric studies found effects at three percent failures or greater, effects were observed for the present studies at two percent or greater, apparently due to the presence of map information on the display. Recommendations for cartographic display design and future research are discussed. Master of Science

When dot-matrix characters are rotated, as might be the case in a moving map display, their dot-matrix patterns are distorted and their legibility is thus affected. In this experiment, 16 subjects performed a random search task, in which they were asked to look for a target in a random character pattern. The independent variables were the direction (clockwise or counterclockwise) and the angle of stimulus image rotation, and the target character's distance from the center of screen, which was also the center of rotation; the dependent variables were response time and response correctness. Significant effects were found in the angle of rotation, the target character's distance from the center, and the target character. The results indicate that (1) no angle-dependent mechanism is involved in performing this task and the angle of rotation influences recognition mainly through the distortion of dot-matrix patterns, (2) the target character's (radial) distance from the center of screen is the determining factor for search time, while the x and y coordinates of the target contributed to dot-matrix pattern distortion, and (3) the target characters interacted differently with the angle and distance factors to determine the extent of distortion and their legibility. Means to quantify the extent of distortion were discussed and the direction for future research is offered. Master of Science

The current study compared the performance of six commonly used cursor devices (absolute touchpad, mouse, trackball, relative touchpad, force joystick, and displacement joystick) on three types of tasks (target acquisition, text editing, and graphics). Prior to these comparisons, each of the devices was optimized for display/control dynamics in independent experiments. A total of 30 subjects were used in the five optimization studies. For each device, the optimization experiment compared a range of control dynamics using a target acquisition task (i.e., positioning a cross-hair cursor over square targets of varying sizes and screen distances). An analysis of variance procedure was used to determine the best control dynamics, of the range studied, for each device. Performance was based on a time-to-target (TT) measure. A comparison of the six optimized devices was then performed on the three task environments. For the target acquisition, text editing, and graphics tasks, a total of 12, seven, and six subjects were required, respectively. For the target acquisition study, the six devices were compared on a task identical to the optimization task; that is, cursor positioning performance for various target sizes and distances. In addition to the TT dependent measure, bipolar scale and subjective rank data were also collected. The text editing task required subjects to perform document correction on the computer using each of the six devices, with cursor keys added as a baseline device. Task completion time (TCT), bipolar scale response, and subjective rank data were collected. For the graphics task, subjects were required to perform basic graphics editing tasks with the six devices. As with the text editing task, TCT, bipolar scale, and rank data were collected. Results indicated a wide variation in the cursor positioning performance of the devices on the three tasks. Without exception, the mouse and trackball performed the best of the six devices, across all tasks. In addition, these devices were most preferred. In general, the two joysticks performed worse on the target acquisition and graphics tasks than the two touchpads. On the text editing task, however, the rate—controlled joysticks performed better than the touchpads. Ph. D.

A broad range of element size, element shape, and interelement spacing/element size ratio combinations at high and low ambient illuminance levels was used to determine optimal combinations for a 5 X 7 dot matrix visual display. Three tasks - a reading test, a random search task, and a structured (menu) search task - were used to enhance the utility of the research. All tasks were displayed on a CRT display 1.02 m from the subject. There were different overall results for the reading and the search tasks, as shown by an analysis of variance. The smaller elements (0.76 and 1.14 mm) and the 0.5 space/size ratio produced the shortest reading times, while the shortest search times came from the larger elements (1.14 and 1.52 rum). The square shape and low illuminance level (5.4 lux) had the shortest times for all three tasks. For a general purpose display, the combination of a square 1.14 rum dot, with 0.57 mm edge-to-edge spacing is recommended, to be used with the lowest practical ambient illuminance. Master of Science

The predictions of a movement control model which states that the degree of motor program control decreases as movement duration increases, were tested in three experiments. The findings indicated that the standard deviations of movement times increases as movement duration increased. However, at longer movement durations, the presence of a visual feedback display decreased the standard deviations of movement times. In Experiment III, the rate of movement was manipulated by varying movement distance while holding movement duration constant. Movement distance did not significantly affect the standard deviations of movement times, which suggests that the degree of motor program control is a function of the duration and not the rate of movement. The effects of movement duration, distance, and the presence of a visual feedback display on the means and standard deviations of start times - the interval from a signal to initiate movement until movement began - were also assessed. Distance did not significantly affect either the means or the standard deviations of start times. However, the shortest movement duration always produced the shortest and least variable start times. Continuous visual feedback increased the standard deviations of start times as compared to a terminal feedback or knowledge of results condition. It was suggested that the presence of continuous visual feedback "interfered" with the initiation of movements. The effects of the independent variables on the ratio of the standard deviations of the two temporal components of the response - SD (start time)/ SD (movement time) - were reported. Master of Science

The amenability of various color-controlling algorithms to the real-time operator control of color stimuli was investigated. Mathematical models based on eight color spaces were employed: three uniform color spaces (L*u*v*, L*h*C*, and Y 2.2u' VI), a graphics algorithm (HLS), an opponent color model, the NTSC broadcast signals (YIQ), and two sets of color primaries. Eighty subjects, divided equally among the color spaces, were required to match colors under time-limited conditions. The apparatus employed was a color-manipulation device using LEDs, custom-built in the Displays and Controls Laboratory at Virginia Tech. The device allows for 12-bit resolution on each color channel and higher stability of luminance and chrominance over the short and long term than can be achieved with experimental quality CRTs. Master of Science

This study used a standardized evaluation to compare two direct manipulation graphics interfaces. The Interfaces investigated were (1) object-oriented (vector) graphics and (2) bit-mapped graphics (object-oriented graphics interfaces are not to be confused with object-oriented programming or object-oriented data bases). Experienced and novice subjects performed objectively derived benchmark tasks appropriate for two-dimensional graphics packages. Both performance and preference data were collected. Task completion time, aborted attempts, learning effects, and errors constituted the performance measures. For the preference data, subjects completed questionnaires to rate the interfaces on both an absolute and a relative basis. Results indicate that the object-oriented graphics interface is superior to the bit-mapped interface. Both the performance and preference measures indicate that the object-oriented interface is better suited for manipulation of graphics. Master of Science