Introduction

The rapid and accurate interpretation of the aircraft detected in a defended area largely determines the success of US Navy air defense (or anti-air warfare - AAW) operations. Commanders are provided with computer-generated graphical representations of detected aircraft as a basis for this interpretation. Rapid and accurate interpretations are dependent on the effectiveness of the representation. With advances in computer processing and graphics capabilities, display hardware technology, and interaction devices, new options for graphical representations can be explored.

One such exploration is being done at the Johns Hopkins University Applied Physics Laboratory where a prototype AAW system that among many other functional capabilities provides a more realistic 3D perspective display of the aircraft, surface ships, and other tactically significant features in the vicinity of a carrier battle group. The use of a perspective display is intended to reduce the commander's situation assessment burden when compared to the traditional plan-view display.

The purpose of this study, performed in conjunction with the Applied Physics Laboratory and the University and Maryland, was to compare human performance on the perspective display with performance on a traditional plan-view display in the accomplishment of a typical AAW task: quickly determining if aircraft in the vicinity of a carrier battle group are ascending, descending, or maintaining altitude. In conventional plan-view displays, where the observer's only available view is from directly above the aircraft, an operator selects or "hooks" an aircraft symbol to obtain amplifying textual readouts for information such as altitude, ascent, and descent. Most existing displays do not directly provide ascent and descent information, but require an operator to monitor altitude changes over time to determine ascending, descending, or steady flight. With a more realistic perspective display, the indication of ascending, descending, or steady flight is integrated directly and naturally into the aircraft representation while more precise information may still be obtained from text readouts.

Background

Although the naval air defense domain is operationally different than traditional air traffic control, the displays used for both show great similarity. There have been some studies that investigated the use of perspective displays for both naval tactical displays and for air traffic control. Among these are a study by Bemis, Leeds, and Winer (1988) which investigated the utility of a perspective naval tactical display, and one by Burnett and Barfield (1991) that analyzed the use of a perspective display for air traffic control. In the Bemis study, the participants were presented with an air traffic scenario of 30 aircraft within a 24-mile area. Using either the plan-view or perspective display, the task was to identify any potentially hostile aircraft that appeared in the display and to assign an interceptor to each identified threat. Their results showed that operators made 34 percent few errors in properly detecting potential threats and 40 percent fewer errors in properly selecting an interceptor for the threat when they used the perspective display. When they analyzed the time required to perform these tasks, the results were less conclusive. For the mean time to detect the presence of a threat, the results for the two types of display were almost equal with a large standard deviation. The perspective display did show a 35 percent advantage in the mean time required to select an interceptor.

Several more studies have investigated the use of perspective displays in place of plan-view displays in the cockpits of commercial or military aircraft. These studies include Ellis, McGreevy, and Hitchcock (1987), and McGreevy and Ellis (1991). In Ellis (1987), ten airline pilots were presented with a perspective display of the airspace around their aircraft. Their task was to monitor the development of a possible collision situation between their aircraft and another aircraft and to recommend an avoidance maneuver. Performance with the perspective display was compared to performance for the same task using a plan-view display. As a measure of performance, average time to react to the collision situation was measured for 6 different angles of approach for the intruding aircraft (from 0 degrees, or a head-on approach, to 150 degrees). The time to react was better by about 8 percent for the perspective display for all approach angles except the head-on case. The authors concluded that the head-on case showed poorer performance because the pilots had difficulty interpreting the perspective display to determine if the intruding aircraft was coming or going. McGreevy (1991) gives a thorough summary of the elements of an effective perspective display for an aircraft cockpit, but its principles could be applied equally well to any type of perspective information display. An interesting conclusion of this study is that a perspective display becomes especially useful only if there is a dynamic spatial relationship between the objects. This dynamism can be provided either by object motion relative to each other or by movement of the observer's point of view.

Still more work has been done in the fundamental theory of perception and interpretation of three-dimensional objects presented on two-dimensional displays. Some of the more notable studies in this area are Attneave and Frost (1969), Attneave (1972), McGreevy, Ratzlaff, and Ellis (1985), McGreevy and Ellis (1986), Grunwald, Ellis, and Smith (1988), Ellis (1990), and Tharp and Ellis (1990). These studies show that observers of a rotated object tend to judge that object's angle of rotation with a mean error that is dependent on the true rotation angle of the object. These results would lead us to conclude that objects viewed from certain slant angles are inherently more difficult to judge than others when viewed on a perspective display. This finding seems to be supported by the results of Ellis (1987) where aircraft viewed from a head-on perspective were more difficult to interpret than those viewed from other angles.

Variables and Hypothesis

The independent variables for our experiment were the display type (plan-view or perspective) and number of descending aircraft (10 or 21 out of a total of 50). The dependent variable was the time required to correctly identify descending aircraft. Our hypothesis was that participants would correctly identify the descending aircraft more quickly when they were using the perspective display than when they used the plan-view display.

Continue

Table of Contents