U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-RD-03-065
Date: September 2004
In-Vehicle Display Icons and Other Information Elements: Volume I
PDF Version (8.33 MB)
PDF files can be viewed with the Acrobat® Reader®
Chapter 4: Icon Recognition
Introduction: Level of realism refers to the relationship between the portrayed object or concept and the graphic means used to represent that object or concept. The graphic realism of an icon can have an impact on comprehension. Recognition can be facilitated for some icons by adding small details, while others are improved by a reduction in detail or by emphasizing important features. Only details that increase symbol recognition and comprehension should be included. Details that detract from recognition and comprehension should be omitted.
Figure 4-1. Five Levels of Realism
Discussion: Reference 1 discusses five different levels of realism that can be used: photographic realism, simplified drawing, caricature, outline, or silhouette. Each of these styles has its place, where one may work better than another for conveying particular types of information. In the driving context, the more detailed styles, such as photographic and simplified drawings, are the least practical. The amount of information they attempt to display makes them less recognizable in the sizes necessary for display. The silhouette style is the simplest of the five styles and is the most common type used in road signs. It is likely that this style will also be the most effective for use in in-vehicle displays.
Design Issues: When icons are being designed as a set, it is important that the same level of detail or realism be chosen to portray the entire set. When this is not possible, the designer should choose both a primary style and an alternate that is most similar (i.e., caricature and outline). This will help to not only increase the ease with which they are recognized but to help the user see them as a set of related icons.
Level of Detail, p. 4-4; Identifying Icons as Part of a Group, p. 5-8
References: 1. Horton, W. K. (1994). The icon book: Visual symbols for computer systems and documentation. New York: J. Wiley & Sons.
Introduction: Level of detail refers to the amount of detail necessary for recognition of a symbol. Only those details that contribute to the meaning of the symbol should be included, while those that distract from the true goals of recognition and comprehension should be omitted. A significant detail refers to a symbol element that would reduce icon recognition and comprehension if removed.
Figure 4-2. Example of Appropriate Level of Detail
Discussion: Icons should be designed with the appropriate level of detail. Including details that are necessary for discriminability can make objects much easier to recognize. However, including excessive amounts of detail simply creates clutter and can contribute to driver distraction. Therefore, care should be taken to omit any unnecessary details from the design.
To ensure that the level of detail is appropriate, references 1 and 2 suggest that the design of symbols should take place on a 20 x 20-unit grid; significant details should be no smaller than one square unit. This simple method can help designers adjust the level of detail of a symbol so that it is less likely to distract the driver or clutter the display.
Design Issues: There are several instances in which it may be necessary to increase the size of significant details within an icon: (1) when an icon is particularly important, such as in the case of warnings of imminent danger; (2) when icons suffer from unavoidable design deficiencies, such as poor color combinations, lack of adequate illumination, or excessive complexity; and (3) when it is important that the icon be noticed against a visually complex background.
Determining the Appropriate Size of Icon Components, p. 3-6; Level of Realism, p. 4-2
Introduction: Perceptual principles of icon design refer to design recommendations based solely on the visual characteristics of the icon without reference to its intended function or meaning (see also references 1 and 2).
Figure 4-3. Perpetual Principles of Icon Design
Discussion: Many of the perceptual theories about how we decode and comprehend symbols have come from the ideas and experiments that were completed by such Gestalt psychologists as Kohler, Wertheimer, and Koffka. They were the first to discover that the determinants of shape and form are the figures in the visual field, which in turn are characterized by their contour (i.e., outline or boundary). This delineation of one part of the visual field from another is called the "figure/ground" phenomenon. The Gestalt psychologists also derived many of the principles that comprise the concept of "figural goodness," whereby the perceptual process of decoding incoming stimuli is enhanced by the inherent clarity and stability of the form (reference 4).
Easterby (references 2, 3, and 4) has conducted multiple studies that examine the figural aspects of symbols (e.g., the lines, curves, and graphics that make up a symbol) and using the principles of "figural goodness" to determine how they affect perception, recognition, understanding, and learning. He argues that the structural properties of a symbol are important determinates of its perceptibility and that they provide the contextual cues that define the meaning of a symbol (reference 1). These structural properties include aspects such as continuity, closure, symmetry, simplicity, and unity (reference 2).
Composition of an Icon, p. 2-10; Design of Prohibition Symbols, p. 4-10
Introduction: Flash rate refers to the rate at which a signal alternates between an illuminated and a non-illuminated state. References 1 and 2 provide several recommendations for the use of flashing signals.
Figure 4-4. Schematic Examples of the Appropriate and Inappropriate Use of Flashing Icons
Discussion: According to reference 2, flash rates should be well below that at which a flashing light appears to be a steady light, which is approximately 30 times per second. Reference 1 recommends rates of about 3 to 10 per second, but states that 4 per second would be best. And reference 3 provides the same recommendation, adding that the flash duration should be at least 0.05 second.
Design Issues: Flashing signals are an excellent means for attracting the attention of the driver; therefore, it is extremely important that they be used only to signal emergency situations. Their ability to quickly gain and divert the driver's attention makes them an unsafe means for presenting noncritical or status information. It is also important to understand how the overuse of flashing signals can actually produce the opposite effect that may be desired. As more than one flashing light is introduced into the environment, the amount of time it takes a driver to react actually increases. Reference 4 found that, if even one irrelevant background light was flashing, reaction time would be greater than for a steady signal. Therefore, it is recommended that backgrounds remain steady when using a flashing signal and that only one signal flash on a display at a time.
Ways to Use Icons, p. 2-6; Conveying Urgency with Icons, p. 5-14
Introduction: Design of prohibition symbols refers to icons that present a specific action and communicate that the action should be avoided. Prohibition should be indicated by overlaying the action symbol with a red circle and 45° slash (top left to bottom right) or cross.
Figure 4-5. Examples of Alternative Methods for Indicating Prohibition
Discussion: A glance legibility study (reference 1) asked participants to match an image they had been shown for either 1/10 s or 1/25 s to various traffic symbols on an answer sheet. Performance was much worse for those symbols with slashes either in front of or behind the pictorial than for those symbols with either a partial slash or no slash. It was concluded that the conventional circle and slash obscured portions of the pictorial and increased the overall complexity of the pictorial. Several recent studies, however, have found that people prefer symbols with the slash either in front of or behind the pictorial (references 1 and 2) and view these symbols as being more effective. Care must be taken when determining the placement of the slash, because a larger symbol may obscure a posterior slash, concealing that the symbol conveys a prohibition. Reference 3 hypothesizes that this may be due to viewers' familiarity with this format, since most symbols are currently being designed this way, or it could be due to the Gestalt principles of good figures.
The effectiveness of prohibition symbols is lowered by ambiguity, complexity, and confusability (reference 3). Ambiguity refers to symbols on which the slash has obscured some detail of the pictorial, making it possible for the symbol to have several different meanings. Complexity refers to symbols that attempt to include too much information (i.e., both an action and its consequence). Confusability refers to symbols that have unfamiliar or unusual features.
Design Issues: There are some ambiguities and inconsistencies across the literature on this topic. A red circle and slash is recommended by several standards documents (see references 4 and 5) for warning pictorial design. This design approach is also consistent with the perceptual principles discussed on page 4-6. However, research suggests that this format may actually obscure the symbol, making the icon much harder to recognize (references 1, 2, and 3). Reference 2 indicates that careful consideration of the placement of the pictorial is necessary. Potential solutions to the problem of obscuring the symbol include changing the orientation of the pictorial, modifying the pictorial so that all important features are displayed or, in some cases, rethinking the overall concept behind the symbol and redesigning it completely. Reference 6 presents designers with several alternative methods for indicating prohibition when the standard 45-degree top-left-to-bottom-right slash covers too much of the symbol.
Composition of an Icon, p. 2-10; Perceptual Principles of Icon Design, p. 4-6
Introduction: General vs. specific icons refers to the degree to which an icon provides information about a general class of in-vehicle messages vs. providing detail about the specific nature of an in-vehicle message. General icons are defined as icons that provide the driver with information about a broad driving situation or class of conditions without specifying detailed information about the situation or conditions. For example, a general icon for the message “crash warning” would indicate that a crash is imminent, but would not convey information regarding the precise nature of the projected crash (e.g., side, front, or rear crash). Specific icons are defined as icons that do provide more detailed information about a driving situation or conditions. For example, a family of specific crash warning icons could be used, with each icon describing the specific nature of the projected crash (e.g., side, front, or rear crash). The table below shows some examples of general and specific icons for key in-vehicle message categories.
Figure 4-6. Examples of General and Specific Icons for Key In-Vehicle Message Categories.
Discussion: Currently, the typical icon development approach entails the assignment of a specific icon to each driver message. For example, specific icons are used to depict low fuel conditions, weather conditions, and motorist services. The number of specific icons presented to drivers increases with the proliferation of In-Vehicle Information System (IVIS) devices. Unfortunately, this will place greater cognitive and memory burdens on the driver and may eliminate the advantages associated with using visual icons in the first place. In particular, working- and long-term memory capacities are quite limited (see reference 1), especially when retrieval strategies (chunking, rehearsal, auditory redundancy) strategies are not available (reference 2).
An alternative approach to using many specific icons to communicate individual messages is to use general icons to convey information about a class of conditions. Such an approach will work best in situations where the driver may not need detailed or specific information to understand the message sufficiently to take the appropriate driving action(s). The key advantage of general icons over specific icons is a reduction in the total number of icons that would be used within the in-vehicle environment.
Reference 3 investigated driver perceptions of the accuracy and acceptability of generic vs. specific icons for a range of in-vehicle message types. The effects of icon type (general vs. specific) were strongly mediated by the scenario descriptions given to subjects. Overall general icons were selected as the most accurate when subjects were presented with a general description of a particular driving scenario, and specific icons were selected as the most accurate when subjects were presented with a specific description of a particular scenario. However, for two message categories-collision avoidance and water recreation-specific icons were selected as the most accurate icons regardless of the scenario description. For the collision avoidance icons at least, this suggests that specific icons are desired where safety is an issue, perhaps because drivers want to have as much information as is available.
These general findings, however, should be considered in light of results from the study relating to the acceptability of general vs. specific icons. While the perceived accuracy of icons varied as a function of the scenario described to the subjects. High levels of acceptability were obtained for both the general and specific icons, regardless of scenario description. Specifically, the general icons resulted in 80 percent or higher levels of acceptance in 23 out of 23 messages in the general scenario description condition and in 20 out of 23 messages in the specific scenario description condition (exceptions were one of the three emergency vehicle messages and both vehicle maintenance messages). Therefore, it seems very clear that general icons are capable of meeting driver expectations and preferences for a broad range of IVIS messages.
Design Issues: Reference 3 only investigated driver acceptance of general vs. specific icons. Driver behavior and performance issues for the "general vs. specific" question remain to be investigated.