Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-99-089
Date: December 1999

Research, Development, and Implementation of Pedestrian Safety Facilities in the United Kingdom


6. Pedestrian Crossings with Signal Control

6.1 Pelican crossings

The main type of independent or stand-alone signal-controlled pedestrian crossing in the United Kingdom is the Pelican crossing which was introduced in 1969. In 1979 local authorities were permitted to install Pelicans without special authorization from the DOT. Major changes were introduced in the 1987 Pelican Crossing Regulations. (DOT, 1991). Since 1979, Pelican crossings have been widely introduced and estimated to be more than 11,000 Pelican crossings in the United Kingdom (Hunt and Lyons, 1997).

The Pelican crossing has a far-side red man/green man signal aspect. Pedestrians register a demand at the push-button and approaching traffic is monitored, usually by microwave detectors. Towards the end of the pedestrian crossing phase, the green man changes from constant to flashing. There is a three-aspect signal head to control vehicles, including a flashing amber phase which permits drivers to go if all pedestrians have cleared the crossing. For pedestrians, however, the red/green man is advisory only; there is no offense of jay walking, or equivalent, in the United Kingdom. Modern Pelican crossings are now Vehicle Actuated, i.e., they monitor approaching vehicles and recall the pedestrian phase more quickly if there are no vehicles approaching. However, substantial numbers of Fixed Time Pelican crossings remain in operation.

As noted earlier, the criteria for deciding if and what type of crossing should be provided is set out in LTN 1/95. The installation of a Pelican (or other type of crossing) will not necessarily reduce pedestrian accidents. It may even result in some increase in pedestrian accidents because of increased pedestrian activity or other factors. Studies have attempted to find relationships between accident rates and levels of pedestrian and vehicle flow. A recent study (CSS, 1997), however, found no correlation.

6.2 Puffin crossings

During the 1990s the DOT has been sponsoring experiments with other types of signal-controlled crossing, particularly the Puffin and the Toucan (below). The Puffin crossing is intended as a replacement for the Pelican.

The need to develop improved pedestrian crossings, and the research program, is described by Billings and Walsh (1991). The Puffin has been developed in response to shortcomings of the Pelican, namely

  • Inadequate time for slow pedestrians to cross.
  • Stressful and confusing nature of flashing green man.
  • Unnecessary delay to vehicles when pedestrians push the button but are able to cross before the green man shows; also when pedestrians complete the crossing early.
  • Excessive delays for pedestrians because of fixed minimum time between pedestrian phases.

At traffic signals with pedestrian phases (but not at Pelicans), there is also the "dead spot" between the green man being extinguished and the red man appearing. (There is no flashing green /flashing amber.) This can cause confusion and anxiety to pedestrians.

Research has been undertaken to develop new technical equipment, and operating strategies, and to assess user behavior. Davies (1992) provides the results of the first Puffin experiments, supported by European Community funds from the DRIVE program. These Puffins used pressure-sensitive mats at the curbside to detect waiting pedestrians and infra-red on-crossing detection to adjust the pedestrian crossing time. Despite technical problems, the results were considered sufficiently positive to continue development work.

Reading undertook further work on Puffins, into user behavior and pedestrian detection. "The stressful conflict between the pedestrians and vehicles during the flashing amber period is eliminated and replaced by stages of clearly defined priority." (Reading, Dickinson, and Barker, 1995). However, the Puffins and consequently some aspects of the experiments, suffered from inadequate and unreliable equipment. For example, 7.6 percent of valid pedestrian requests were undetected and a number of false cancellations occurred. The potential for using computer vision-based pedestrian detection systems is explored (Reading, Wan, and Dickinson, 1995) as these would (potentially) allow detection of not only the presence but also the volume of pedestrians. However, it was concluded that the vision-based systems available at the time were too restrictive for the Puffin task and that long-term development would be necessary.

Crabtree (1997) describes tests that combine the TRL signal control operating system MOVA (Micro­processor Optimised Vehicle Actuation) (DOT, 1997) and VSPD (Volume Sensitive Pedestrian Detection) with the Puffin (and Pelican) crossing. Two forms of VSPD were used: both used computer processed video images. "The most significant outputs from the project were the VSPD which can sense the numbers of pedestrians waiting to cross and a version of MOVA that can take input from the detector and include it in its optimization process." In terms of delays, MOVA with or without VSPD performed better for both vehicles and pedestrians than VA. The VSPD MOVA with Puffin was more responsive to pedestrian demand but gave greater vehicle delays compared to the MOVA Puffin. At the Puffins, some safety behavior changes were noted. More pedestrians were looking at the traffic rather than straight ahead (where the green man would be located on the Pelican). There were fewer serious crossing infringements, i.e., fewer pedestrians crossed during the green to traffic, "probably because of the reduced delay for all Puffin crossings." However, there were more slight infringements, i.e., pedestrians crossing within the red to traffic but outside the green man period.

Further tests have been carried out, and there are now over 60 Puffin test sites. Despite problems with the reliability of the equipment, the DETR is sufficiently convinced that the Puffin should replace the Pelican. The combination of Puffin and MOVA and possibly VSPD makes the Puffin responsive to local conditions and gives the potential to adjust priorities to suit local policies. Regulations enabling local authorities to install Puffin crossings without Government approval came into force in December 1997. Performance specifications have been issued to industry, and it is believed that suitable on-crossing detection systems and pedestrian demand units (incorporating the push-button and near-side signal) will be produced in the near future. (The performance specification does not stipulate the type of technology.) A Pelican costs approximately £10-15,000 ($16,000 - $25,000) at 1997 prices and a Puffin costs approximately £2,000-3,000 ($3000 - $5000) more than that.

6.3 Toucan crossings

As cyclists are not permitted to cycle across Zebra or Pelican crossings, the Toucan crossing is designed for shared use by pedestrians and cyclists. During the 1980s, the DOT developed a parallel signal-controlled crossing for pedestrians and cyclists. However, this proved to have limited applications as it was expensive, required considerable curb-side space and was generally considered to be an overkill. Trevelyan and Ginger (1989) found that where cyclists cycled over Zebra or Pelican crossings, there were no safety or practical problems for pedestrians. As a result, a shared crossing — the Toucan — was developed (Morgan, 1993). This has a red man/green man and a green bicycle aspect on a single far-side pole. It is push-button operated often with additional vehicle actuation for pedal cycles. There are now over 200 Toucan crossings in the United Kingdom although they still require special authorization from the DETR. More recent installations include infra-red on-crossing detection and nearside aspects, like the Puffin. Further research has been undertaken by TRL for the DETR into various technical specifications and user behavior (Taylor and Halliday, 1997). As with the Puffins, there have been problems with the reliability of the equipment but user response (from pedestrians and cyclists) has been favorable. It looks likely that future versions of the Toucan will be very similar to the Puffin but with cycle aspects too.

6.4 Further developments in signal-controlled pedestrian crossings

Hunt and Lyons (1997) and others have been critical of the operating strategies for signal-controlled pedestrian crossings (including those at traffic signal junctions) because of the delays to pedestrians and the consequent risks that arise from crossing outside the green man period. They suggests various improvements to reduce pedestrian delays, including:

Radical approach: At sites where pedestrian activity is high and vehicle occupants are to be given low priority, the signals should default to red-to-vehicles rather than (at present) red-to­pedestrians

Balanced approach: Allowing the pedestrian precedence period to start when there are five vehicles or less in the detection zones 25-150 m (82 - 492 ft) upstream. Model simulations of this approach suggest lower cycle times, reduced pedestrian delay and crossing during red man, and small increases in vehicle delay.

6.5 Signal control at junctions

It is now usual to include pedestrian crossing facilities within busy signal-controlled junctions. Where traffic signals have been introduced to control traffic at roundabouts (usually caused by capacity problems), these have also proved beneficial for pedestrians.


United States Department of Transportation - Federal Highway Administration