The article below is a draft submitted to the Transport Research Arena 2016 conference.


Traffic lights in urban transport systems are often designed to take into account the situation on the road, i.e. detect vehicles and optimize road traffic flow. A wide variety of methods is employed to detect vehicles - from cameras through induction loops to laser radars. However, pedestrians are usually left with the rather antiquated method of pressing a button.

Such an approach can generate a number of problems, especially if the light settings are based on the premise that manual and automatic activation of lights are equally effective, i.e. that every pedestrian will by default press the button at the moment s/he approaches the crossing. Such an assumption (one of several made when installing pedestrian buttons) is false due to a number of factors. These include the need for the pedestrian to: identify that the lights need to be activated (which is the case at only selected crossings), find a push button, walk up to it and activate it. Another assumption is that the pedestrian will wait until the light changes to cross the road, regardless of the situation on the intersection. While this has been shown to be a false premise in the past, it is particularly unlikely to hold true when the pedestrian sees that parallel traffic streams have a green light while s/he has a red one (due to not pressing a button or pressing it too late). Additional problems may stem from inadequate accessibility of the buttons, the insufficient number or noticeability thereof, or their being situated off the default (shortest) path.

While pedestrian buttons are often treated as a default solution, manual detection is in fact a way of reducing the attractiveness and effectiveness of walking as a mode of transport. This may not be obvious, but it is enough to imagine a driver forced to go through the process delineated above (having to identify whether the lights will not change by themselves, if so - looking around for a way of activating them, moving to a given point where the lights can be activated, waiting for them to change, then going back to the track which was leading from the source of his/her journey to the destination) - at every intersection - to see that they are in fact a problem.

The paper presents a discussion of approaches to the activation of pedestrian lights based on the situation in Poland, with particular focus on Warsaw, where field research was conducted. The problems outlined above are quantified and possible solutions thereto are suggested.

Keywords: pedestrian crossings; pedestrian detection; push buttons; sustainable urban transport; Warsaw;

1. Introduction

Broadly understood Intelligent Transportation Systems are lately the vogue, particularly in urban areas, where they are sometimes presented as the remedy to traffic problems. Paradoxically, these supposedly state-of-the-art solutions often include manual activation of pedestrian traffic lights (push buttons). In effect, higher investment and upkeep costs may result in worsening rather than improvement of pedestrian traffic conditions. These costs are – at least in Poland – usually externalized, i.e. not taken into account when assessing the effectiveness of the system. Implementing user-activated traffic lights may in effect have a negative impact on the sustainability of a given transport system by discouraging not only pedestrian journeys, but also intermodal ones with pedestrian elements (i.e. ones including public transport).

The paper will begin with a short review of Polish legal regulations concerning detection of pedestrians and vehicles at intersections. Next, the possible advantages and disadvantages of manual pedestrian detection will be discussed. Apart from theoretical arguments, data will be provided from field research conducted in Warsaw on several intersections equipped with push buttons. In conclusion, some suggestions will be made as to a more rational and effective approach to pedestrian detection.

2. Polish regulations concerning pedestrian detection

Detection of pedestrians and vehicles is regulated in Poland by a regulation of the Minister of Infrastructure on road signs, traffic lights and road safety devices (Minister Infrastruktury (2003)). The regulation has been updated several times during the last twelve years, but the parts concerning detection of pedestrians have remained unchanged. In part 3 of Attachment 3 two categories of detection devices (detectors and buttons) are listed in the definition of traffic lights, and later a more detailed definition of a detector is provided. Interestingly, despite the initial differentiation between (automatic) detectors and (manual) push buttons, the latter are later included among the former even though they are activated manually (i.e. do not detect anything themselves).

The broader definition of detectors includes various types of automatic detection (inductive loop, magnetic, infrared, laser, magnetic resonance, microwave, radiowave, video, proximity sensors). Other, similar solutions are also allowed. Manual detection, on the other hand, is limited to push buttons. Although manual detection of vehicles is forbidden, automatic detection of pedestrians is outright allowed (but not obligatory). Further on, more detailed requirements concerning detectors are specified. A telling detail is that vehicle detectors are required to be “reliable […] and easy to install and use” while no such qualitative requirements are specified in the case of push buttons. The latter need to be located at a given height above ground level, in positions identified according to an analysis of pedestrian traffic flow. They should have a durable, visible and safe casing and be able to confirm being pushed by a sound or visual signal. At first sight and in theory such requirements should ensure an effective system of pedestrian detection. Whether this is true in practice will be discussed later on in the paper.

Apart from technical requirements concerning detectors, the regulation lists the desired effects of installing traffic lights, among them the improvement of pedestrian traffic conditions. Other desired effects include improving traffic conditions for public transport and/or minor roads and coordinating traffic lights at different intersections. The regulation underlines that more than one of these effects should be achieved. However, the required elements of a traffic light cost-to-gain analysis do not take pedestrians into account at all. The only costs which need to be taken into account at this point are those borne by the road authority (installing the traffic lights) and vehicles (delays, faster wear, collisions resulting from a vehicle running into another one stopping at a red light).

The regulation points out that pedestrians and other road users may ignore red lights if they are not effectively managed. In effect, signals automatically accommodated to actual traffic conditions are strongly recommended. (As mentioned above, the regulation indicates that push buttons do not qualify among the elements of an automatically accommodating system.)

Further on, among detailed rules concerning signal lights, the regulation underlines that traffic lights should be treated as a last resort if other solutions cannot be implemented. The analysis of available alternatives should take into account pedestrian traffic. Pedestrian-related situations in which traffic lights may be justified include: high traffic intensity resulting in a long waiting time for a chance to safely cross the road, a high number or risk of accidents involving pedestrians, a high volume of pedestrian traffic and “the presence of children and the disabled among pedestrians, especially on higher-speed roads”. These criteria (except for the last one) are further on quantified.

To sum up, the regulation on traffic lights does not directly recommend the use of push buttons. Traffic lights are to be used only if safety and traffic conditions cannot be otherwise improved and a wide variety of automatic modes of detection is allowed. The fact that automaticity of accommodation is underlined suggests that manual detection should rather be avoided (effective automatic pedestrian detection was already available in 2003 when the regulation was adopted – cf. SRF (2003)). Notwithstanding such regulations, push buttons are usually the default choice. (For a long time they were even the default mode of detection of bicycles on separated bicycle paths, even though the regulation expressly obliges that vehicles be detected automatically.) Such an approach may be partly due to the relative simplicity of installing manual detection (especially when the requirements concerning its location are ignored). Another reason, however, appears to be a narrow understanding of traffic conditions as limited to vehicular traffic. Such an approach does not take into account the impact of signalization on pedestrian traffic conditions or treats them superficially. (This is visible not only in the types and locations of pedestrian traffic detectors but also in the traffic light sequences.)

3. The advantages and disadvantages of manual pedestrian detection

As mentioned above, Polish regulations indicate that traffic lights should serve to improve pedestrian traffic conditions. Lights can also serve to improve traffic conditions for other users, but pedestrians should not be clearly disadvantaged, as this may lead to their ignoring the lights. The non-discrimination of pedestrians is also an integral element of a sustainable urban transport system which promotes walking, cycling and public transport while minimizing the demand for car use.

Manual pedestrian detection may theoretically serve to improve the travel conditions of pedestrians. However, it also entails several negative consequences. The impact of manual detection on the following aspects will be analyzed:

• Pedestrians’ waiting time for green light,

• Road safety / risk of traffic light violation,

• Cost-to-gain ratio,

• Accessibility for the disabled,

• Efficiency and competitiveness (attractiveness) of walking.

The analysis will include references to the field study of pedestrian behavior at intersections with manual pedestrian detection. The study was conducted during one week in September 2015 at eight intersections in Warsaw, twice at each intersection, and is described in detail in the appendix.

3.1. Pedestrians’ waiting time for green light

Manual pedestrian detection may theoretically shorten the time pedestrians wait for a green light in comparison to the situation on a crossing without pedestrian detection, i.e. with green lights in every cycle. In order for such an improvement to take place, two conditions need to be met. First, pressing the button must result in a shortening of the cycle so that the pedestrians receive green earlier than they would in a regular cycle. Second, the gain resulting from that shortening must be greater than the time necessary for the pedestrian to identify the need to press a button, locate it, possibly walk up to it if it isn’t located on the pedestrian’s default route, press it, and possibly walk balk to his/her original route.

Hypothetically, if every pedestrian wanting to cross the street immediately pressed the button upon arriving at an intersection and pressing the button would always result in a green light appearing earlier than the parallel green for vehicles would appear in a cycle without pedestrians, push buttons would improve pedestrian traffic conditions. Such a situation (the shortening of waiting times for pedestrians) is especially desirable in urban areas which according to current European and Polish urban policies are to be designed and organized so as to encourage the choice of sustainable modes of transport (cf. EC (2013), MIR (2015)). A sustainable approach to urban transport includes optimizing traffic lights so as to minimize waiting times for the non-motorized.

In practice, however, the use of push buttons is largely at odds with such goals, especially if pressing them serves as a condition of the pedestrian green appearing at all. An international study for the European Commission indicated that “the installation of push button traffic lights only intensified the problem by extending the waiting times for pedestrians even more” (Monheim, Frankenreiter (2000): 36). The Warsaw field study has confirmed push buttons are having the same negative impact on pedestrian waiting time in Warsaw fifteen years later. This is the result of neither of the above-mentioned conditions being met: (1) the definite majority of pedestrians does not push the buttons and (2) the traffic lights are not programmed so as to quicken the pedestrian green when the button is pushed (at least in the daytime), but only condition it appearing at all.

Of the 2058 pedestrians observed, only 19% immediately pressed the push button when arriving at a crossing equipped therewith, and 4% did so with some delay. In other words, nearly 78% of those arriving at a crossing did not push the button at all. This is a significantly higher figure than that indicated in similar studies conducted in the United States in the 80s and 90s (cf. Zegeer et al. (1983), Palamarthy et al. (1994) – however even in these studies, only half of the pedestrians pushed the buttons) and similar to the results of a study in the early 2000s in which 68% of pedestrians did not push the button (cf. Campbell et al. (2004): 83). (It should be noted that even placing the push button in a place where pedestrians have to look – i.e. on the same panel as the pedestrian traffic lights at Puffin crossings – did not ensure that it would be pressed: “up to 49% of pedestrians crossed without using the signal demand button” (Walker et al. (2005): 63).) In the Warsaw field study, some of the pedestrians did not press the button due to somebody else having done this earlier, but in fact most of the pedestrians either did not notice or ignored the push buttons altogether. This is visible in the fact that on average only 1.4 pedestrians pushed the button per crossing per cycle even though there were on average 6 pedestrians at each intersection with an average of nearly 8 buttons at their disposal. In 6 out of 8 cases there were on average fewer pedestrians than pushbuttons. The assumption that at least one pedestrian arriving at each extreme point of the crossing will push a button turned out to be false. More often than not, people were able to cross only thanks to someone at another part of the intersection having pressed a button.

Pushing the button did not hasten the green light, nor did it always result in a green light in the same cycle, even if the parallel road traffic received one. In effect, there was no possibility of pedestrians receiving a green light sooner than they would if the buttons were not there. The light could only be activated at the moment it would appear without buttons or not be activated at all, i.e. the walking conditions could be the same as they would be without the buttons or worse.

Theoretically, a situation could take place in which no parallel road traffic appeared, and there was no parallel green light for vehicles. In such a case, a pedestrian pushing the button would probably activate the green light faster than it would be activated by parallel vehicles. However, no such situation was observed. It is more likely to take place during nighttime. During daytime the buttons were shown to enable solely the worsening of traffic conditions for pedestrians by lengthening their waiting time.

3.2. Road safety / risk of traffic light violation

The prolonging of pedestrians’ waiting time for a green light can result in their deciding not to cross the street at all (choosing a different route), continuing to wait, or crossing on a red light. None of these situations contribute to the sustainability of a transport system, but the last one can be particularly troublesome in terms of road safety. Pedestrians on an intersection with push buttons are likely to see that cars on the parallel road have a green light, wait a little longer, assuming that the pedestrian light is delayed, and finally decide to cross the street despite the red for pedestrians. By the time they decide to do this, they are much less likely to cross the whole street before the perpendicular light changes to green.

“One of the most important things we discovered…is that people will only wait so long [to cross a street] and then they’ll take a risk,” said Ron Van Houten, a professor of Western Michigan University, adding that “uncertainty -- about whether the push button works (…) reduces pedestrian waiting compliance” (Roadway Safety Institute (2014)). A positive correlation between longer waiting times for pedestrians and dangerous pedestrian behavior at signalized intersections has also been confirmed by research elsewhere (cf. Brosseau et al. (2013), Monheim, Frankenreiter (2000): 25, Jamroz (2014): 130). Interestingly, the problem is ignored by both the police and road administrators in Poland. If taken up at all, the inverse is suggested: that any elongation of red lights for cars will be dangerous by provoking drivers not to wait for a green light. (NB: Drivers in Poland relatively often ignore the beginning of a red light, but situations in which they do not wait for it to end are extremely rare.) Such a purely hypothetical situation is presented as a rational argument for installing traffic light programs which empirically result in pedestrians disobeying red lights (cf. (2015)). The field study showed that light sequences requiring pedestrians to wait for longer than one cycle in order to cross the street resulted in jaywalking. Taking into account that – as shown above – push buttons at intersections serve primarily to elongate pedestrians’ waiting time, they can only have a negative impact on road safety by increasing the risk of traffic lights violations (cf. Fig. 2).

The arguments presented by the spokesman of the Warsaw Road Authority (risk of red light running due to a longer red for vehicles) are not only unconfirmed in practice or theory, but in fact experiments have shown the inverse to be true: a dwell-on-red setting for cars reduced the risk of accidents (Archer et al. (2008), cf. Monheim, Frankenreiter (2000): 36). Although the research concentrated on evening and nighttime traffic, the authors pointed out that implementing such a system during the entire day could positively influence road safety. A similar solution was implemented in Łódź in 2013 at a crossing between a church and supermarket on Retkińska street. The default light is currently red for vehicles and green for pedestrians. The need for pedestrians to activate the lights via pushbuttons was replaced by vehicles activating them via automatic detectors. In effect, pedestrians do not need to wait as long as before.

3.3. Cost-to-gain ratio

The default practice in Poland, including Warsaw, is calculating traffic light sequences solely from the point of view of car traffic. The intensity of pedestrian traffic may be taken into account in terms of width of pedestrian crossing, but the length and frequency of the green for pedestrians is usually an after-effect of vehicle-oriented calculations. The situation may change as a result of outside interventions (e.g. from NGOs or the media), but by default pedestrians are not taken into account in terms of gains and losses resulting from implementing a given light sequence. They are to make do with the results of vehicle-oriented calculations. (Interestingly, this sometimes entails a non-optimal light sequence for vehicles as well due to excessive spare capacity – cf. Hughes (2010).)

The same approach is adopted when calculating road throughput, which is usually - implicitly or explicitly - presented by traffic engineers as the overarching goal of road infrastructure. Unfortunately, road throughput is by default calculated solely for vehicles and in number of vehicles. The number of people who can get through is not taken into account (i.e. by calculating that one bus may be the equivalent of 50-100 cars) and pedestrians are ignored altogether. In effect, road safety and the sustainability of the transport system are treated as secondary goals. The popularity of pushbuttons in Polish urban areas is a symptom of such an approach.

The field study showed that pedestrians were present at 88% of the cycles at all intersections. The figure was lower at only one site and in three there were pedestrians during 98-100% of cycles. The average number of cycles in which pedestrians received a green light, however, was over 10 p.p. lower (less than 78%). At half the sites, over 12% of cycles were ones in which pedestrians were present but did not receive a green light. On KEN Ave. this was the case in 26% of cycles. In all these cases, the parallel car traffic streams received a green light. The turning on of a pedestrian green would either have no effect on the car traffic light cycle or lengthen it by a few seconds (the difference resulting from the clearing speeds of pedestrians and vehicles, mitigated by the fact that pedestrians have a shorter distance to cross before leaving the intersection). The pedestrians, on the other hand, were required to wait an additional 60-150 seconds (assuming that they pressed the button – those who did not needed up to 6 minutes to get to the other side of the road). The hypothetical gains for vehicles were therefore incommensurate with the losses of the pedestrians.

The only intersection where there was a green light in every cycle in which a pedestrian was present was Wilanowska/Nowoursynowska where the pushbuttons appeared to be turned off (i.e. pressing them was not necessary to receive a green light). The pedestrian green light appeared during 100% of cycles and pedestrians made use of them in 100% of cycles.

3.4. Accessibility for people with reduced mobility

Push buttons are a form of discrimination of all pedestrians. Discrimination is defined as “the practice of unfairly treating a person or group of people differently from other people or groups of people” (Merriam-Webster (2015)). At first sight, push buttons may seem to be the equivalent of automatic forms of vehicle detection. However, as has been mentioned, they require that the pedestrian identify the need to press a button, locate it, possibly walk up to it if it isn’t located on the pedestrian’s default route, press it, and possibly walk balk to his/her original route. The significance of these differences is easier to imagine if the pedestrian’s situation is transposed onto that of the driver by imagining what intersections would look like if vehicles were detected in a similar way to pedestrians. The minimum width of nearly all pedestrian crossings in Poland is 4m (only recently has the legislation changed to allow narrower crossings). This is the equivalent of roughly four “lanes” of individual pedestrian traffic. Even if we optimistically assume that push buttons are located on both sides of the crossing and are adjacent thereto, that still means that a pedestrian has to walk up to one edge of the crossing to activate the lights, regardless of whether this point is on his/her route. An analogical situation has been illustrated in Figure 1. The driver would need to move his/her vehicle to one of the edges of the road, activate the detector, then move back to his/her intended route. It would not be possible to activate the lights from either of the middle lanes or by simply stopping in the middle of the extreme lanes (the wheels would need to be right next to the curb).

Fig. 1. Car detection loops (yellow) situated analogically to pedestrian manual detection.

It should also be remembered that the majority of intersections in Warsaw do not have push buttons on both sides of crossings (cf. Wierciński (2014)), i.e. we should cross out one of the detectors in Figure 1, e.g. the one at the top.

The car detector analogy is especially suitable to illustrate the situation of people with reduced mobility, i.e. the disabled, the elderly or mothers with baby carriages. Like automobiles, they have problems with moving back and forth along the edge of the intersection (i.e. from a point on their default route to a push button and back) and it costs them time and effort. Push buttons located on traffic signal poles are often situated next to the sloping part of the sidewalk leading to street level. The effort required to maneuver a wheelchair to such a position, stop it on the inclination and push the button is incommensurate with the effort required of drivers to activate a traffic light.

3.5. Efficiency and competitiveness (attractiveness) of walking

Push buttons reduce the efficiency and attractiveness of walking for the same reasons that they are a form of discrimination: pushing them requires adjusting one’s route and behavior to the push buttons (cf. Fig. 1 & 2).

Fig. 2a. Waiting for red light on shortest path (green line) from source to goal (push buttons marked in yellow)

Fig. 2b. Natural behavior after light fails to change to green

The inverse is the case with traffic lights for vehicles: the lights are supposed to accommodate to the situation on the road. In the case of pedestrians, the users are expected to perform a sequence of actions in order to activate the lights. In effect, longer routes need to be taken, attention cannot be focused on other things (e.g. talking on the phone, talking to a friend), additional light cycles need to be waited through, etc. Meanwhile, the driver needs only observe the traffic lights and react thereto. In effect, some of the natural advantages of walking (e.g. the possibility of taking the shortest possible route) are lost and it becomes a less attractive form of transport in comparison to the less sustainable automobile. Public transport also loses competitiveness, as it is nearly always accessed by foot.

4. Conclusions

The most obvious conclusion is that manual pedestrian detection should be discontinued altogether as a necessary condition of activating traffic lights, especially in urban areas. As has been shown, push buttons used as a sine qua non make pedestrian traffic less competitive and attractive, more prone to accidents, less efficient and often entail higher losses than gains. Their widespread use in Poland is symptomatic of a strictly car-oriented approach to traffic engineering which ignores the needs and behavioral patterns of pedestrians (cf. Monheim, Frankenreiter (2000): 36) and entails further repressions rather than remedies to the problems it generates. Even though the problematic nature of such an approach was already noticed in the 1980s (cf. Baass 1989), a typical “solution” consists of sending (municipal) policemen to catch jaywalkers while ignoring the cause of the problem. Such an approach is also at odds with the Polish legal regulations concerning traffic lights which require that jaywalking should be prevented by not disadvantaging pedestrians.

Pedestrians should automatically receive a green light if a parallel traffic stream does. This is the optimum solution for pedestrians because it also provides green for those who arrive at the crossing right before the light would change or after it has changed to green. The field study showed that this group accounted for 35% of all pedestrians (7% arrived right before the lights changing, and 26% during the green phase).

In places and/or times where/when parallel traffic streams do not appear regularly, pedestrians should either receive a green light by default or be detected automatically. The first option is particularly suitable during evening and night hours when road traffic is less intense and more prone to excessive speeds. A default green for pedestrians should also be used on local roads and in areas with high volumes of pedestrians/where pedestrians should be prioritized (i.e. city centers). The second option (green light activated by automatic detection) is suitable in places where pedestrians rarely appear and road traffic is intense. Automatic detection has been shown to be at least equally effective to manual detection with infrared technology achieving the best results (>90%) in terms of both sensitivity and selectivity (cf. Markowitz et al. (2012). The detection of 90% of pedestrians would be a fourfold better result than the percentage of pedestrians who pressed pushbuttons in the field study.

Automatic detection should encompass not only pedestrian waiting zones, but also the crossings themselves so that not only the frequency, but also the length of the green/perpendicular red light is adapted to the situation on the road. Such a solution is already being implemented in the United Kingdom on Puffin crossings (Department for Transport (2006): 11, 40). The use of automatic detection can thus be beneficial not only to pedestrians (by ensuring that they receive a signal and that it allows them to cross the road), but also to drivers (by shortening the pedestrian green/perpendicular red light if conditions allow).

Implementing the suggestions listed above would contribute to the improvement of pedestrian traffic conditions and encourage a more sustainable model of mobility, both in terms of congestion and health/safety/pollution.

Appendix A. Detailed results of field study in Warsaw

The field study consisted of observing and recording pedestrian behavior at 8 intersections in Warsaw, twice at each intersection, from 22 to 28 September 2015. Table 1 presents the results of the study in percentage points. In Table 2 the same results are presented in numbers. The detailed parameters of the places and times analyzed are presented in Table 3.

Table 1. Results of field study (percentages).

Street crossed (x intersection) cycles with pedestrians cycles with green light for pedestrians pedestrians pressing button (pedestrians pressing button) immediately (pedestrians pressing button) after a delay pedestrians not pressing button pedestrians arriving during red light (pedestrians arriving during red) when lights changing pedestrians arriving during green light
Dobra (x Lipowa) 100% 86% 19% 16% 3% 81% 70% 9% 30%
Żelazna (x Twarda) 93% 80% 34% 24% 10% 66% 79% 8% 21%
Koszykowa (x Krzywickiego) 44% 37% 48% 42% 6% 52% 73% 9% 27%
Konstytucji Sq. (southern part) 91% 78% 22% 20% 3% 78% 70% 2% 30%
KEN Ave. (x Ciszewskiego) 98% 72% 25% 21% 4% 75% 81% 10% 19%
Dolina Służewiecka (x Nowoursynowska) 100% 100% 30% 25% 5% 70% 68% 7% 32%
Sikorskiego Ave. (x Wilanowska) 88% 84% 21% 18% 3% 79% 77% 11% 23%
Witosa/Becka (x Czerniakowska) 87% 79% 14% 11% 2% 86% 74% 2% 26%
Average 88% 78% 22% 19% 4% 78% 74% 7% 26%

Table 2 Results of field study (numbers).

Street crossed (x intersection) cycles with pedestrians cycles with green light for pedestrians pedestrians pressing button (pedestrians pressing button) immediately (pedestrians pressing button) after a delay pedestrians not pressing button pedestrians arriving during red light (pedestrians arriving during red) when lights changing pedestrians arriving during green light
Dobra (x Lipowa) 50 43 62 52 10 265 228 29 99
Żelazna (x Twarda) 43 37 52 37 15 107 125 12 34
Koszykowa (x Krzywickiego) 19 16 16 14 2 17 24 3 9
Konstytucji Sq. (southern part) 42 36 42 37 5 147 133 4 56
KEN Ave. (x Ciszewskiego) 49 36 73 61 12 218 235 28 56
Dolina Służewiecka (x Nowoursynowska) 50 50 74 62 12 176 169 18 81
Sikorskiego Ave. (x Wilanowska) 44 42 92 79 13 342 336 47 98
Witosa/Becka (x Czerniakowska) 33 30 51 42 9 324 279 6 96
Total 330 290 462 384 78 1596 1529 147 529

Table 3. Parameters of crossings analyzed

Street crossed (x intersection) road classes* Urban zone** cycles observed persons per cycle (avg) inter-section arms observed number of buttons on crossings observed days of observation time of day
Dobra (x Lipowa) L/L I 50 7 2 4 23.09 15:40-16:00, 16:04-16:28
Żelazna (x Twarda) Z/L I 46 3 2 4 22.09 09:45-11:15, 18:20-18:40
Koszykowa (x Krzywickiego) Z/L I 43 1 1 2 22.09 09:10-09:40, 17:50-18:15
Konstytucji Sq. (southern part) Z/Z I 46 4 1 2 27.09, 28.09 19:20-20:00, 15:58-16:28
KEN Ave. (x Ciszewskiego) Z/Z I/II 50 6 2 6 25.09, 28.09 15:40-16:07, 12:10-12:35
Dolina Służewiecka (x Nowoursynowska) GP/Z II 50 5 2*** 20 25.09, 28.09 16:00-16:55, 12:48-13:30
Sikorskiego Ave. (x Wilanowska) GP/G II 50 9 2 10 26.09, 28.09 17:00-17:45, 12:48-13:30
Witosa/Becka (x Czerniakowska) GP/GP II 38 10 2*** 15 26.09, 28.09 12:10-13:10, 14:35-15:40

* L – local road, Z – collector road, G – main road, GP – fast traffic main road, (S – highway, A – freeway)

** I – city center area, I/II – border of city center and urban area, II – urban area, (III – suburban area)

*** There were also two right-turn slip lanes at Dolina Służewiecka and one at Becka – each one was equipped with traffic lights and push buttons.


Archer, J., Candappa, N., Corben, F., 2008. Effectiveness of the Dwell-on-Red Signal Treatment to Improve Pedestrian Safety during High-Alcohol Hours, 2008 Australasian Road Safety Research, Policing and Education Conference, November 2008.

Baass, K.G., 1989. Review of European and North American Practice of Pedestrian Signal Timing. Proceedings of the 1989 Annual Conference of the Roads and Transportation Association of Canada, Calgary, Alberta, September 17-21, 1989.

Brosseau, M., Zangenehpour, S., Saunier, N., Mirando-Moreno, L, 2013. The Impact of Waiting Time and Other Factors on Dangerous Pedestrian Crossings and Violations at Signalized Intersections: a Case Study in Montreal, Transportation Research Part F Traffic Psychology and Behaviour 09/2013.

Campbell, B.J., Zegeer, C.V, Huang, H.H., Cynecki, M.J., 2004. A Review of Pedestrian Safety Research in the United States and Abroad, US Department of Transportation, January 2004.

Department for Transport, 2006. Puffin Crossings. Good Practice Guide – Release 1,

European Commission (EC), 2013, Annex: A Concept For Sustainable Urban Mobility Plans to the: Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions: Together Towards Competitive and Resource-Efficient Urban Mobility, COM(2013) 913 final, Brussels, Dec. 2013.

Hughes, T., 2010. Recent New Zealand research on walkable environments, presentation, November 2010.

Jamroz, K. (Ed.), 2014. Ochrona pieszych. Podręcznik dla organizatorów ruchu pieszego, Krajowa Rada Bezpieczeństwa Ruchu Drogowego, Gdańsk, Kraków, Warszawa.

Markowitz, F., Montufar, J., Lovejoy, K., 2011. Automated Pedestrian Detection: Challenges and Opportunities, Walk21 Conference, Vancouver, October 2011.

Merriam-Webster Dictionary, 2015. „discrimination”,

Minister Infrastruktury, 2003. Rozporządzenie Ministra Infrastruktury z dnia 3 lipca 2003 r. w sprawie szczegółowych warunków technicznych dla znaków i sygnałów drogowych oraz urządzeń bezpieczeństwa ruchu drogowego i warunków ich umieszczania na drogach (Dziennik Ustaw nr 220, poz. 2181 z dnia 23 grudnia 2003 r.)

Ministerstwo Infrastruktury i Rozwoju (MIR), 2015. Krajowa Polityka Miejska 2023. Projekt., Warsaw, Sept. 2015.

Monheim, H., Frankenreiter, G., 2000. Town and Infrastructure Planning for Safety and Urban Quality for Pedestrians: State-of –the-art Report,, European Commission, Luxembourg.

Palamarthy, S., Mahmassani, H.S. , Machemehl, R.B., 1994. Models of Pedestrian Crossing Behavior at Signalized Intersections, Report 1296-1, Center for Transportation Research, University of Texas at Austin, January 1994. Cited in: Campbell, B.J., Zegeer, C.V, Huang, H.H., Cynecki, M.J., 2004. A Review of Pedestrian Safety Research in the United States and Abroad, US Department of Transportation, January 2004.

Roadway Safety Institute, 2014. “Advancing pedestrian safety through countermeasures that work”, in Roadway Safety Institute News, Vol. 1, No. 2, Summer 2014,

SRF Consulting Group, Inc., 2003. Bicycle and Pedestrian Detection: Final Report. Prepared for United States Department of Transportation, Federal Highway Administration., 2015. “Przejścia tylko dla pieszych-sprinterów”,

Walker, R., Winnett, M., Martin, A., Kennedy, J., 2005. Puffin crossing operation and behavior study, PPR239 Published Project Report, TRL Limited, August 2005.

Wierciński, P., 2014. Letter to the Ombudsman concerning pedestrian detection, [zobacz >>>].

Zegeer C, Opiela K, Cynecki M., 1983. Pedestrian Signalization Alternatives. Report No. FHWA/RD-83-102, Federal Highway Administration. Washington, DC, 1983. Cited in: Safe routes, Why are pedestrian push buttons used at traffic signals?,