Fifteenth International Technical Conference of the Enhanced Safety
of Vehicles (ESV)
Melbourne, Australia, May 1996
Paper Number 96-S11-W-22
Published by NHTSA, Washington,
USA.
PDF version.(47K) Project Report (160K)
A dilemma is that bright warning lights might cause discomfort and glare at dusk or at night. The authors examined the geometry of a typical scenario for a car encountering a bus at the side of the road. It was found that a warning light system could be specified which achieved the required signal range but which, due to its high mounting position on the bus and sharp vertical cut-off of the light distribution downwards, enabled motorists to move into a lower intensity portion of the beam as they approached the bus.
In New South Wales (NSW) current practice is to have "wig wag" yellow flashing lights and signs at the front and rear of the bus. The authors were engaged by the NSW Department of Transport to investigate the suitability of several proposed signalling systems, together with the current system. This paper describes the outcome of an analysis of the technical and visual ergonomic requirements of signalling systems fitted to school buses.
To be effective the system must satisfy each of three requirements (after Lay, 1981):
The motorist will require a distance away to see the signal (the signal range) which takes into account the distance travelled during the response time to the signal, the distance travelled during slowing down to 40km/h and the distance over which to stop from 40km/h, if necessary (the buffer zone). These three stages are illustrated in Figure 1.
Figure 1. Derivation of signal range for vehicles approaching the
bus
from either direction.
It is preferable that the motorist does not brake heavily because this may be a hazard to following traffic and it could also lead to reluctance to slow down if school buses with lights flashing are repeatedly encountered on the road. On a level road at 100km/h a typical vehicle will decelerate at between 0.5 and 1 metres per second per second (m/s/s) without the use of brakes. Under gentle braking a deceleration of 2m/s/s is regarded as comfortable. Heavy braking involves decelerations of around 5m/s/s (all decelerations described in this paper are average, not peak).
From these values the distance at which a signalling system on the bus has to be first seen by an approaching motorist can be calculated. The formula is:
s = ((V*V - v*v)/2a) + Vt + d . . . . . (1.)
Table 1 shows the application of Equation 1 to several scenarios.
Table 1 - Signal Range (m) Required to Slow from 100km/h to 40km/h,
Including a 30m Buffer Zone
Type of braking | Deceleration m/s/s | Distance for typical reaction time (2.5s) | Distance for alert reaction time (1.5s) |
None (engine braking) | 1.0 | 424 | 396 |
Gentle | 2.0 | 261 | 234 |
Heavy | 5.0 | 164 | 136 |
On the basis of this analysis, the signal on a school bus should be visible and recognisable at no less than 250m for buses operating in 100km/h areas (this assumes some gentle braking will be required and it includes a 30m buffer). A minimum of 100m is required for buses operating in 60km/h areas but a common signal range of 250m for all buses is preferable for uniformity.
The human eye is more sensitive to a light source the closer that source is to the line of sight. This means that the further a signal is from the line of sight the brighter it will need to be to elicit a response. The necessary luminous intensity of a signal will also increase as the square of the distance away. The relationship is:
I = 2Kd*dL/1,000,000 . . . . . . .(2.)
where
I = Optimum luminous intensity of a steady red signal for a required
signal range (cd)
K = (a/3) to the power 1.33
a = angle of the signal from line of sight (degrees,minimum 1 degree)
d = required signal range (m)
L= background brightness (cd/sq m)
The formula is the outcome of considerable research (Cole & Brown, 1968, Fisher & Cole, 1974). The optimum intensity is that which invokes, essentially, 100% probability of seeing, coupled with a near minimum reaction time. This and other data form the basis of Australian Standard AS2144 (AS 1989) and international recommendations (CIE 1988) on the photometric specification for traffic signals.
It should be noted that the intensity is directly proportional to the brightness of the background to the signal. Typical values of background luminance range from 10,000 cd/sq m on a bright day to 100 cd/sq m or less around dusk. Therefore the range of a signal of given intensity can vary by a factor of more than 10 depending on background lighting conditions (intensity is proportional to the square of range). This is why signals of relatively low intensity can appear quite adequate for long distances under favourable (dull) lighting conditions while being unsuitable for bright conditions.
Figure 2. Derivation of offset distance - typical car/bus geometry - looking from rear, right-hand drive vehicles
Figure 3. Relationship between signal intensity & signal range for a steady red light, based on constant offset and a background brightness of 10,000cd/sq m.
The intensity values for yellow signals need to be three times that for red for equal visual performance (Fisher & Cole, 1974). This will not normally be a problem in practice since a yellow lens can transmit about 3 times the light from an incandescent lamp over that for a red lens and therefore the wattage of the lamps needed will be the same. Using Equation 2, the following signal intensity requirements can be deduced, as shown in Table 2.
Table 2. - Signal Intensities for Two Signal Ranges with Steady Signals Viewed Against a Sky Background of 10,000cd/m2 and an Offset of 5.5m
Signal Range (m) | Steady Red Signal (cd) | Steady Yellow Signal (cd) |
100 | 210 | 630 |
250 | 390 | 1170 |
Table 3 - AS2144 Minimum Traffic Signal Intensity (cd)
Type of signal | Range (m) | Red Intensity(cd) | Yellow Intensity(cd) |
General Purpose | 100 | 200 | 600 |
Extended range | 240 | 600 | 1800 |
The values at 100m signal range match those of Table 2. The values for extended range traffic signals exceed those of Table 2 since traffic signals have a greater offset (Hulscher 1975).
In the case of yellow aftermarket lamps intended for use as either front or rear vehicle turn signals, a manufacturer would logically aim for an intensity around 200cd which is the maximum permitted for single-intensity rear mounted turn signals. In bright daylight these would provide a signal range of about 100m when used on a car or small trailer but they become ineffective when high-mounted on a large vehicle such as a bus. Referring to Figure 3, a 200cd yellow signal is equivalent to a 67cd red signal which has a range of about 100m when used on a car but less than 25m when high-mounted on a bus.
Table 4. - SAE J887 Guidelines for Yellow Signal Units (cd)
- - - - - - - - - - - Degrees left (right is mirror image)
Degrees up/down | 30 | 20 | 10 | 5 | 0 |
10 | - | - | - | 50 | 125 |
5 | - | 375 | 750 | 750 | 750 |
0 | 75 | 450 | 1000 | 1250 | 1500 |
-5 | 75 | 500 | 750 | 1125 | 1125 |
10 | - | - | - | 100 | 100 |
Flashing Signals - Lights may be made to flash. Contrary to popular belief, a flashing light is more difficult to detect initially than a steady one of the same intensity. However, once detected a flashing light is more likely to demand inquiry or be taken notice of than a steady light. In order to maintain the same signal range, the intensity of a flashing light will need to be increased over that of a steady light (Cole 1972, Holmes 1971).
Assuming a signal to flash at 60 cycles per minute (typical for flashing turn signals), with the off time equal to the on time, then the intensity will need to be increased by a factor of about 1.4 times to that derived from Equation 2. Even greater intensity would be required for a faster rate of flashing but, in any case, there are technical limits to the rate at which automotive lamps can be flashed (e.g. losses due to incomplete heating up of the filament and decreased service life).
Cycle Time -A property related to flashing rate is cycle time. The message should become unambiguous when a complete cycle of the signal system has elapsed and the next cycle begins. In the case of single colour wig-wag signals this will be after three light operations (e.g. left, right then left) and the total time will be about 1.5 seconds. This duration is similar to the response time used in Equation 1. In effect the cycle time is part of the process of recognition of the signal.
Dirt & Deterioration - The signal may become dirty and the hardware deteriorate over time. Taking these factors into account a nominal factor of 1.1 is used to cover the in-service deterioration of signal intensity.
Backboards - In the case of buses it is not practical to fit a black backboard of sufficient size to improve the signal range. This is because the angle between the edge of the signal and the outer edge of the backboard needs to be about 1o in order to effectively isolate the signal from the background (Fisher & Cole 1974). This translates to a backboard diameter of about 4m when viewed from 250m. No allowance is therefore made for backboards.
Derived signal intensities - Applying these factors to the values in Table 2, and rounding the results, leads the values in Table 5.
Table 5. - Necessary Signal Intensities for Flashing Signals on School Buses
Signal Range (m) | Red (cd) | Yellow (cd) |
250 | 600 | 1800 |
100 | 300 | 900 |
50 | 200 | 600 |
The limitation of these adverse effects is generally given much attention in the provision of lighting and signalling at night. Light directly towards the eyes of motorists is kept to the minimum practicable.
The specified maximum intensity limit for red traffic signals is 1000cd (AS 1989), there being no limit to the yellow signal "in view of the relatively short intervals for which such signals are normally displayed". The standard suggests that the 1000cd should normally satisfactorily limit glare, at night, from signals used on roads where traffic route lighting is installed. However where roads have local road lighting or are unlit authorities are advised to consider installing signals with intensities not greater than 350cd.
Essentially similar values of intensity are embodied in road lighting standards (AS 1973) viz, 1000cd and 500cd maximum intensities for the light emitted at the horizontal from luminaires used for traffic route and local road lighting respectively.
In Australia the maximum values for vehicle yellow turn signals are 700cd and 200cd for front and rear signals respectively. The maximum intensity from the white low beam headlight in the direction of oncoming motorists is 437.5cd.
There is evidence that the intensity of a yellow light can be higher than that of a white light before being deemed unsatisfactory; the results of some investigations suggest that it can be 40% greater (van Bommel & de Boer 1980).
Practice leads to the conclusion that, in order for a light not to be glaring when viewed at night, it should have a maximum intensity of about 1000cd in the direction of view, preferably less if the road is poorly lit or unlit. Reference to Table 5 shows that the required peak intensity of the light beam of a yellow bus signal (1800cd) needs to be greater than this in order to fulfil its alerting role.
A third way is to give careful attention to the light beam shape. This is done for the vehicle headlight low beam in which the high intensity portion for forward seeing and the low intensity portion for limiting glare are sharply separated. The authors therefore examined whether this approach could be applied to the school bus signal.
Only at 250m away is it necessary for the motorist to experience the elevated intensity (1800cd), whilst closer to the bus (100m) the required signal intensity decreases substantially (900cd), even though still producing a clear signal. When very close to the bus (50m and less) the motorist should only be subjected to the same intensity as would be experienced with conventional turn signals, i.e 200cd to 700cd.
In order to set out a specification for the complete angular light intensity distribution for a signal light it is necessary to analyse the angular position of the signal in the field of view as the motorist approaches the bus. A desirable outcome is that motorists are in the high-intensity part of the beam some distance from the bus in order to be alerted and then move into a lower intensity portion of the beam when they get closer to the bus, to alleviate any potential over brightness of the signal.
Using the offsets of the motorist to the bus shown in figure 2. (viz eyes to far signal light; 2.2m vertical and 5.0m horizontal), the angular offsets during approach to a bus are obtained as shown in Table 6.
Table 6. - Angular Offsets for Various Distances from the Bus and the Required Signal Intensity for a Yellow Flashing Signal
Distance Away d (m) | Horizontal Offset (degrees) | Vertical Offset (degrees) | Required Signal Intensity (cd) |
250 | 1.2 | 0.5 | 1800 min |
100 | 2.8 | 1.3 | 900 min |
50 | 5.7 | 2.5 | 600 min |
25 | 11.3 | 5.0 | (600 max) |
12.5 | 21.8 | 10.0 | 200 max |
To cope with some vertical misalignment of the signal the maximum intensity requirements at d=25m. (av=5 degrees) should be 600cd; this value will provide the required signal intensity but will also restrict potential over-brightness.
The values of the required intensities for 50m and beyond are minimum values. These need to be associated with maximum values to avoid the potential for excessive brightness of the signal. Taking into account available technology a maximum intensity not more than 1.5 times the minimum intensity was applied. The minimum values in Table 6 have therefore been reduced by half this tolerance (i.e. by 25%). Thus the maximum values will be only 25% above the values in Table 6. The adjusted minimum values will result in only a 10% reduction in signal range, whilst providing a tolerance in design and manufacture. In practice manufacturers are likely to design signal lights well within the tolerance range and the resulting intensities are likely to be close to those given in Table 5. A model specification based on these considerations can be constructed and is given in Table 7.
Table 7 - Recommended Intensities for a Flashing Yellow Signal Light
(cd)
- - - - - - - - - - Degrees left (right is mirror image)
Degrees Vertical | 30 | 15 | 10 | 2.5 | 0 |
5 (up) | - | - | - | - | 500 |
3 | - | - | - | - | 700 |
1.5 | - | - | - | - | 1400 |
0 | - | 500 | 700 | 1400 | 1400 |
-1.5 | - | - | - | - | 1400 |
-3 | - | - | - | - | 700 |
-5 | - | - | - | - | 500 |
-10 (down) | (200) | (200) | (200) | (200) | (200) |
Australian Standard AS2144 (1989) "Traffic Signal Lanterns". Standards Australia.
CIE (1988) "Guide for the Design of Road Traffic Lights", Publication 79. Commission Internationale de l'Eclairage (Vienna).
CIE (1975) "Colours of Light Signals", Publication 2.2. Commission Internationale de l'Eclairage (Vienna).
Cole, B.L. and Brown. B.(1978) "A Specification of Road Traffic Signal Light Intensity". Human Factors 10.(3).
Cole, B.L. (1972) "Visual Aspects of Road Engineering". Proceedings of Australian Road Research Board 6.(1).
Fisher, A.J. and Cole, B.L.(1974) "The Photometric Requirements of Vehicular Traffic Signal Lantern". Proceedings Australian Road Research Board. 7(5).
Fisher, A.J. and Hall, R.R. (1972) "The Influence of Car Front Design on Pedestrian Accident Trauma". Accident Analysis and Prevention 4, pp47-58.
Holmes, J.G. (1971) "The Language of Flashing Lights" in "The Perception and Application of Flashing Lights". Adam Hilger Ltd (London).
Hulscher, F.R. (1975) "Photometric Requirements for Long Range Road Traffic Light Signals". Australian Road Research 7(5).
Lay, M. (1981) "Source Book for Australian Roads". Australian Road Research Board
RTA (1992) "Safety of School Children Near Buses". Road Safety Bureau, Roads and Traffic Authority (NSW)
SAE J887 (1987) "School Bus Warning Lamps". Society of Automotive Engineers (Warrendale).
Staysafe (1994) "Pedestrian Safety 1. School Children Around Buses". Parliament of NSW Standing Committee on Road Safety.
Van Bommel W.J.M and de Boer J.B. (1980) "Road Lighting". Kluwer Technische Boeken B.V. (Antwerp).
Under the Australian Design Rules emergency vehicles are required to have flashng warning lamps with a daylight range of at least 200m "in any direction". School buses and emergency breakdown vehicles mostly use the warning lamps when they are stationary and, in any case, they do not have "right of way" over other vehicles. Therefore, a high intensity signal to the sides is unnecessary. Indeed, at night, a high intensity beam directed towards the sides can reflect off shop windows and wet surfaces and cause confusion about the actual location of the hazard.
In recent years there has been a trend with Australian emergency service vehicles, such as Police vehicles, to use an array of flashing, fixed beam lamps instead of rotating beacons. We examined the visual ergonomics of these lamps, compared with rotating beacons.
On-axis signal intensity for a yellow signal
to achieve a signal range of 250m in bright daylight.
Type of signal | Intensity (cd) |
Steady yellow | 1300 |
Flashing fixed-beam yellow | 1800 |
Rotating beacon (yellow) | 9100 |
As described in the above paper, in order to avoid excessive glare at night, a yellow light should have an intensity of no more than 1000cd. The typical flashing fixed-beam lamp had twice this intensity and the typical rotating beacon had fourteen times this intensity. A further disadvantage is that motorists are less likely to move under the high-intensity portion of the beam of a rotating beacon as they approach the scene, compared with a fixed-beam lamp (see the analysis for school bus lights).
In the case of emergency breakdown vehicles, the most hazardous situations
are likely to arise at night, particularly when it is wet. In these circumstances
the rotating beacons, which must have a very high intensity to achieve
a satisfactory daytime signal range, are likely to be counter-productive.
Approaching motorists could be prevented from having a clear view of the
scene of the breakdown due to the unavoidable glare from the beacons.