Click for next page ( 7


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 6
6 Report Organization tion they expect to see an object of importance. Cox (2) points out that warning-light placement faces the following constraints This report documents the work performed in this project. in terms of capturing attention: Chapter 1 describes the research approach and summarizes the findings of the literature review. Chapters 2, 3, and 4 describe Motorists make multiple decisions on proximal events while the photometric characterization, the static screening, and driving. As a result, distant low-effective-intensity lights may the performance experiments, respectively. Chapter 5 pro- not be detected since they have no immediate interest. vides conclusions and suggested research. A more in-depth Drivers' eyes are cast downward in the natural human pos- discussion of the experiments is included as appendixes to this ture. As a result, drivers do not naturally keep a special report (not published herein but available on the TRB website lookout for distant objects. at www.trb.org/news/blurb_detail.asp?id=9632). A meaningful proportion of motorists is colorblind or has Trade or manufacturers' names appear in the report solely poor visual acuity. because they are considered essential to the object of this report. Conspicuity Literature Review The conspicuity of an object, which refers to how well it Currently, there is a wide disparity in the design of warning captures one's attention, comes in two types: attention con- lights used on roadway maintenance vehicles. The Manual on spicuity and search conspicuity. Attention conspicuity is attrib- Uniform Traffic Control Devices (MUTCD) provides some uted to an object's characteristics, such as proximity, color, and guidance for the use of traffic control, signage, and auxiliary movement. Search conspicuity refers to an object's ability to safety vehicles, but these recommendations are not specific as to the nature of the lighting on the vehicles themselves. capture one's attention when one is actively searching for Recent studies have highlighted the lack of specific criteria for it, such as a retroreflective street sign placed in a consistent the use of warning lights. Kamyab and McDonald (1) found location (3). that there is a great disparity among states in the use of warning Because the presence of a maintenance vehicle is unexpected, lights and that disparity exists even within local agencies. vehicles that have poor attention conspicuity are less likely to Several studies have also investigated different aspects of be detected at a safe distance because motorists are not actively lighting systems, from color to configuration, but no single searching for them. study has been undertaken to provide comprehensive guide- Detection and recognition of unexpected events have further lines for the marking of maintenance vehicles. Because many implications than simply attention conspicuity. The next sec- types of maintenance vehicles are required to operate in bad tion explains why response time to unexpected events is longer weather environments and during both daytime and night- than that to expected events. time, these environments are particularly critical aspects of the warning-light guidelines. A more detailed literature review is Reacting to the Unexpected provided in Appendix A. Whenever individuals respond to events that have been perceived, they are transmitting information. Information Hazard Detection and Recognition is defined by Shannon and Weaver (4) as the reduction of Hazard detection refers to the first stage of information pro- uncertainty. Information is potentially available in an event cessing in which an object is perceived by one's senses. Hazard any time there is uncertainty about what the event will be. recognition refers to a later stage of information processing Information theory states that the uncertainty of an event in which drivers use their memories to relate the object to is dictated by the number of possible events that can occur, previous experiences. Recognition typically involves mental the probability of each event occurring, and the context or operations (or attention) and as a result takes longer than de- sequential constraints that relate each event together (5). tection. However, certain characteristics of objects can support Dewar and Olson (6) explain that motorists operate with pre-attentive processing, or recognition without the applica- a set of expectancies, predisposing them to believe that things tion of attention. will happen in a certain way. There is an increase in driver Contrast sensitivity is a main determinant of one's ability to perception-response time when the expectancies are violated; detect objects of interest in a visual scene because the human this increase can lead to increased driver errors and crashes. eye emphasizes regions of differences in illumination--because To account for these predispositions, advanced warning signs they possess the most information. are used with the intention of establishing expectancy for Detection and recognition of objects in the road is context upcoming hazardous conditions. However, when confronted dependent. Drivers scan the roadway by looking in the direc- with unexpected events, motorists require assistance in coping

OCR for page 6
7 with the change in task demands. A form of guidance, termed Holmes also suggests that people need to be educated on how positive guidance, has been developed to aid people in such to recognize flashing signals (because they are artificial) and situations. how to interpret their meanings. The characteristics affecting warning-light conspicuity in- clude contrast brightness, flash effective intensity, flash color, Positive Guidance flash frequency, flash duration, flash shape, flash type, flash pat- Positive guidance (6) is a way of providing information to tern, flash size, number of elements, and apparent motion, and allow the driver to detect a hazard in a roadway environment steady-burn light color. that may be visually cluttered, recognize its threat potential, select an appropriate speed and path, and complete the required Contrast Brightness maneuver safely. The positive guidance concept acknowledges three levels of driver performance: control, guidance, and Contrast brightness refers to the direct comparison of one navigation (6). reflecting surface with another. Contrast brightness of a flash- ing light signal is obtained from the difference in illumination Control. The control level encompasses the interaction between the lamp-illuminated bulb, called a "roundel," and the between the driver and the vehicle. Drivers control vehicles background. through the steering wheel, accelerator, and brake. Guidance. The guidance level describes the selection and Flash Effective Intensity maintenance of a safe speed and path. Dewar and Olson (6) Roufs (8) defines flash threshold as the minimum effective state that information at the guidance level comes from the intensity increment required for perceiving the flash. For short highway, traffic, and traffic control devices. flashes, the threshold is driven by the product of the flash effec- Navigation. The navigation level refers to the planning tive intensity and duration. The threshold for long flashes is and execution of a trip from origin to destination. Decisions at mainly determined by the effective intensity. the navigation level are made at select points based on infor- mation extrapolated from maps, verbal directions, experience, Flash Color guide signs, and landmarks. Cook et al. (9) investigated the conspicuity of warning bea- cons according to flash color and found that when effective Information Placement intensity is held constant, amber has the poorest detection Dewar and Olson (6) also explain that, to execute positive time under both day and night conditions. Blue light mini- guidance in traffic control devices, four principles of informa- mizes the effects of disability glare and daytime discomfort tion placement must be followed: primacy, spreading, coding, glare. Green light has the quickest detection time during day and redundancy. Primacy requires information on signs to conditions, but is the poorest for disability glare and dis- be placed according to importance to the driver. Spreading comfort glare. Red light yields the quickest detection times requires information content to be spread out across multiple and gives rise to the least discomfort glare. signs when its content is too great to place on one sign. Coding requires pieces of information to be organized into larger Flash Frequency units (e.g., using specific colors and shapes for street signs). Redundancy requires information to be presented in more than Misinterpreting flashing lights designed to communicate a one way at the same time (e.g., an emergency vehicle's visual message to motorists can be as dangerous as missing the sig- warning lights and auditory siren). nal. To avoid misinterpretation, the flashing light signal must be seen for the duration of one period. Holmes (7) states that the flashing signal should be repetitive and have a maximum Flashing Warning Lights interval of 5 s to continuously retain the observer's attention. Flashes are bursts of light which, by definition, are un- expected because they do not occur in nature (save for light- Flash Duration ning). This characteristic is their most important feature and why they are so good at capturing attention. Holmes (7) sug- The flash duration is defined as the length of time during gests that flashing lights have their own language. The flash's which the light is on in one flash cycle. Brown and Gibbs (10) characteristics (such as flash frequency, effective intensity, found that, as the flash frequency duration decreased for fre- and duration) are elements of a language that can be learned. quencies in the range of 1.5 to 3 Hz, there was a corresponding

OCR for page 6
8 decrease in reaction times. However, for signals with frequen- of both a closed-field and field experiment, the following cies of 1 Hz and 0.33 Hz, Gerathewohl (11) found that longer conclusions were made: flash durations yielded shorter reaction times. If only one type of light is used, four-way flashers pro- vide the most accurate information about closure rate and Flash Shape service vehicle speed. The flash shape refers to the temporal distribution of light Adding more of the same type of lights does not increase in the flash cycle. Howard and Finch (12) state that for flashes the amount of information provided to the driver or en- that last longer than the critical duration of 50 ms, the square hance the driver's ability to extract information from the wave pattern is more effective than a triangle shape wave of lights. equal flash energy. Changing the location of the light(s) does not increase the information or ability to extract information; it is important that the light can be seen from all directions. Flash Type Lighting parameters had little effect on driver response. Cook et al. (9) investigated the conspicuity of warning Adding a four-way flasher to any other warning light in- beacons according to flash type. They found that strobe warn- creases the amount of information provided to the driver, ing beacons were subjectively considered to convey greater and combining a roof-mounted flasher light and rotating urgency, while rotating warning beacons were considered to light increases the information to the driver. be less annoying and minimized the effects of disability glare. Apparent Motion Flash Pattern Under certain conditions, it is possible to create a sense Cook et al. (9) also investigated the conspicuity of warning of motion between two stationary sources of light by flashing beacons according to flash pattern. They found that when the two lights on and off with one source temporally trailing more than one warning beacon was present on a vehicle, bea- the other. Foster (15) showed that a model developed to de- cons that flashed simultaneously were detected significantly scribe certain real-motion effects also translated to describe the faster than beacons that flashed alternately. Simultaneously existence of an apparent-motion effect. flashing beacons were also subjectively rated as more con- spicuous, while those that flashed alternately had the lowest Steady-Burn Light Color discomfort glare. Color is an established coding dimension for inter-vehicle signaling. Projector et al. (16), however, reject the use of Signal Size color-coding owing to variation in observer vision, desatura- Many investigations on steady lights at threshold levels have tion of colors in haze and fog, and variation in filter efficien- concluded that lamp size does not play a significant role in cies, but note that color is useful as a redundant perceptual determining its conspicuity. However, the perception of light dimension. under road conditions is quite different than under labora- tory conditions. Cole and Brown (13) concluded that effective Hazard Analysis intensity is independent of signal size for light signals with a high probability of being seen (called optimum signal lumi- In investigating the effective conspicuity of new warning nance) and that, if the lamp is of optimum luminance, its size lights, factors that present potential drawbacks must also be does not matter. considered. Disability glare, discomfort glare, distraction, and eleptogenic response are such factors. Number of Lights Disability Glare Cook et al. (9) investigated the conspicuity of warning beacons according to the number of elements utilized. From Disability glare occurs when a bright light source impairs subjective ratings, they found that the greater the number of an individual's ability to see objects. The effect of disability warning beacons, the greater the perceived conspicuity. glare caused by warning beacons, as stated by Cook et al. (9), In 1990, Hanscom and Pain (14) developed guidelines for was assessed by subjects' ability to detect a pedestrian in their warning-light systems on service vehicles engaged in short- vicinity. They found that disability glare was worsened by term or moving maintenance operations. Based on the results amber beacons, strobe beacons, and maximum intensities.

OCR for page 6
9 Discomfort Glare Weather Discomfort glare is defined as glare that is annoying or The presence of snow, rain, or fog interferes with the per- painful, but that does not cause impairment in the visual field. centage of light reaching the driver's eyes from the warning Discomfort glare could potentially have safety implications light. When the brightness is decreased, the signal is harder because it may cause drivers to avert their gazes. Cook et al. (9) to detect. found that discomfort glare was worsened by amber and green beacons, strobe beacons, maximum flash frequencies, and Road Geometry simultaneous flash frequencies. Because the human eye's sensitivity to light is dependent on location of the light within the retina, it is possible that the Distraction placement of the warning light in the visual scene as a result of A balance needs to be made between warning-beacon con- the road geometry will have an effect on its conspicuity. The spicuity and warning-beacon distraction. Cook et al. (9) found geometry of the road, combined with the obstruction of lights that the presence of a warning beacon is significantly more from trees, rocks, and buildings, may affect the conspicuity of distracting than no warning beacon at all, but the extent of the warning lights. the distraction was not related to flash type, frequency, or effective intensity. State Practices for Roadway Warning Lights The application of warning lights to maintenance vehicles Eleptogenic Response (Epileptic Seizure) differs among highway agencies in the Untied States. Warning- Some features of flashing lights, such as flashing light fre- light specifications for some state departments of transporta- quency, luminance, field of view, and flash type, are relevant tion are presented in this section. These differences highlight to eleptogenic response. Frequencies above 5 Hz should be the need for developing guidelines that will have nationwide avoided. Luminance as low as 20 cd/m2 can trigger eleptogenic applicability. response; however, this exceeds the luminance required to make a warning beacon conspicuous. Lights flashing in the cen- Virginia ter of the visual field are more likely to cause an eleptogenic re- sponse. Also, drivers of emergency vehicles reported that strobe The Virginia Work Area Protection Manual (19) specifies the beacons cause more visual discomfort than rotating beacons. design and application of temporary traffic control devices. The manual states that warning lights should be either a ro- tating amber light or high-effective-intensity amber strobe Environment Complexity light, and that rotating lights shall be mounted to be view- Hargroves (17) states that the background has a significant able for 360 among other specifications of intensity, flash effect on the conspicuity of flashing lights. Day, night, glare, and frequency, etc. irrelevant lights can affect the conspicuity of the flashing-lights signals. Ohio The Ohio Department of Transportation (ODOT) (20) Number of Irrelevant Lights established a vehicle warning-light policy to assure the districts Crawford (18) found that response times to light increase and Central Office maintain uniform lighting array, equipment from 0.8 s to almost 2 s when 21 lights are added to an other- light, marking, and conspicuity. This policy states that all safety wise clear background in a dark soundproof room. He also lighting will be flashing lights; amber in color; composed of showed that steady signals are always more effective than flash- photo strobes, LEDs, or a combination of both; and viewable ing ones if the proportion of flashing background lights exceeds from 360. 1 in 10, and therefore, overuse of flashing lights would defeat their purpose. New York The New York State Department of Transportation Time of Day (NYSDOT) follows a vehicle marking and lighting standard For a fixed luminance, a warning light will have a further that was developed in the mid-1980s. Few recommendations detection distance during the night than it will during the day have been made to improve upon the standard; one change that because the contrast of the signal is great at night. has been implemented is the use of more LED lights for tail

OCR for page 6
10 lights and marker lights to reduce power draw and increase on daytime driving, but could improve the visibility of vehi- visibility. The DOT believes that the halogen rotating yellow cles during nighttime or low-visibility winter weather (21). beacon provides the best overall light for visibility and safety Morgan (29) found that retroreflective tape reduced side and for the traveling public. rear impacts into trailers in dark conditions. Maine SAE Standards The Maine DOT (MDOT) does not have a traffic engineer- The Society of Automotive Engineers (SAE) released a stan- ing handbook, but amber lights are used on all of their con- dard for the lighting and marking of industrial equipment on struction vehicles; state law precludes the use of red or blue. highways (J99) in March 1999 (30). The standard states that there shall be at least two amber flashing warning lamps spaced as laterally wide as practicable and mounted at the same level Iowa at least 42 in. high as measured from the lamp's axis. The Iowa DOT conducted an investigation on the types of SAE also developed SAE J2040 (31) in 2002 to specify the crashes involving snowplows and concluded that the rear end requirements for tail lights placed on vehicles of widths of of the snowplow needs to be more visible to give approaching 2032 mm or wider. The standard states that the color of the vehicles more time to respond. The snowplows currently use tail light shall be red and should have an effective projected two amber rotating beacons and two amber rear-directional luminous lighted lens area of at least 75 cm2. alternate flashing strobes (21). Retroreflective tape, warning flags, and auxiliary headlamps are also used as warning devices. Survey of Current Equipment Available There are many products on the market with similar photo- Texas metric characteristics, and the information provided to con- Texas DOT adopted a warning-light policy for use on spec- sumers is often confusing. Three technologies of flashing ified vehicles and equipment based on research conducted at equipment are available. The difference among these tech- the Texas Transportation Institute (TTI) in 1998 (22). Amber nologies is the source of the light: incandescent filament bulb, warning lights are used to identify highway maintenance and xenon or high intensity discharge (HID) flash tube (commonly service equipment. referred to as strobe lights), and LED. However, because no classification system currently exists, the lighting must be judged on the source technology only. Flashing Light Measurement Issues In the 19th century, it was recognized that intermittent Incandescent Filament Bulb light, or flashing lights, produced higher visibility than a steady There are two types of flashing lights that use incandescent light of the same intensity. Thus, efforts began to quantify the filament bulbs: rotating beacons and 360 flashing lights. The visibility as effective intensity. Effective intensity is defined pulse width and shape of a rotating beacon are determined by as the luminous intensity of a fixed (steady) light, of the same the reflective optic because the bulb is on continuously. The relative spectral distribution as the flashing light, that would wattage of the bulbs describes the quantity of light available have the same luminous range as the flashing light under iden- to the system, but the shape and efficiency of the reflector is tical conditions of observation (23). A singular equation has what controls the pulse intensity and width. The 360 flashing not been developed, but several options for calculation exist, in- light ramps the current up and down quickly to create the time cluding the Allard (24), modified Allard (25), Blondel-Rey (26), dependence. A dome with a Fresnel lens encases the lamp and Blondel-Rey-Douglas (27), and Form Factor (28) methods. focuses the light in the plane of the observer. Retroreflective Tape Xenon or HID Flashtubes During the winter months, detection and recognition of The xenon or HID flashtube lamps have a similar structure snowplows can be deterred by the snow cloud produced by to that of the 360 flashing lights with a Fresnel dome encas- these vehicles. Because the cloud of snow covers the tail light ing the flashtube light source. A significant difference between and makes detection of such vehicles even harder at night, the flashtube and incandescent filament sources is the peak the use of retroreflective strips has been considered. A study instantaneous intensity and pulse width. The peak instanta- conducted by TTI found that the 8-inch-wide orange and neous intensity can be 1000 times more, but the pulse width fluorescent-orange magnetic strips had an insignificant impact is usually 1000 times less in width.