National Academies Press: OpenBook

Traffic Signal Control Strategies for Pedestrians and Bicyclists (2022)

Chapter: Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays

« Previous: Chapter 6 - Treatments that Reduce or Eliminate Conflicts with Turning Traffic
Page 85
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 85
Page 86
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 86
Page 87
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 87
Page 88
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 88
Page 89
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 89
Page 90
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 90
Page 91
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 91
Page 92
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 92
Page 93
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 93
Page 94
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 94
Page 95
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 95
Page 96
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 96
Page 97
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 97
Page 98
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 98
Page 99
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 99
Page 100
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 100
Page 101
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 101
Page 102
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 102
Page 103
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 103
Page 104
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 104
Page 105
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 105
Page 106
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 106
Page 107
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 107
Page 108
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 108
Page 109
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 109
Page 110
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 110
Page 111
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 111
Page 112
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 112
Page 113
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 113
Page 114
Suggested Citation:"Chapter 7 - Treatments that Reduce Pedestrian and Bicycle Delays." National Academies of Sciences, Engineering, and Medicine. 2022. Traffic Signal Control Strategies for Pedestrians and Bicyclists. Washington, DC: The National Academies Press. doi: 10.17226/26491.
×
Page 114

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

85   This chapter describes six treatments aimed at reducing delay for pedestrians and bicycles and accommodating slower pedestrians, grouped by their primary function: C H A P T E R 7 Treatments that Reduce Pedestrian and Bicycle Delays Primary Function Section Treatment Name Reduce effective red time 7.1 Short Cycle Length 7.2 Reservice Increase effective green time for pedestrians and bicycles 7.3 Maximizing Walk Interval Length 7.4 Pedestrian Clearance Settings for Better Serving Slower Pedestrians Limit impact to other road users while serving pedestrians 7.5 Pedestrian Recall versus Actuation 7.6 Pedestrian Hybrid Beacons Additional treatments for reducing delay are found in Chapter 9 (treatments focused solely on cyclist delay) and Chapter 10 (treatments for multistage crossings). The first pair of treatments, Short Cycle Length (Section 7.1) and Reservice (Section 7.2), aims to reduce delay by shortening pedestrians’ and bicycles’ effective red time. “Effective red time” is that part of the signal cycle in which a user group (pedestrians or bicycles) is not intended to begin crossing. In general, delay for all users is minimized when cycle lengths are as short as possible—that is, just long enough to provide capacity for all movements plus additional slack to account for variability. There is one large exception to this rule, however, which is vehicular traffic along a coordinated arterial. For those vehicles, cycle length matters little if vehicles arrive during a green period as part of a “green wave.” To a large extent, traffic signal timing in the U.S. has focused on arterial coordination, which is easier to provide with long cycles. However, long cycles result in long delays for pedestrians, bicyclists, and others (e.g., transit riders, cross-street traffic, and left-turning vehicles) who are not part of the green wave. In many situations, sacri- ficing coordination for shorter cycles can lead to little or no additional delay for vehicles, while substantially reducing delay for pedestrians and bicyclists. “Reservice” means serving a movement—in this case, a pedestrian or bicycle crossing—twice (or more) in a cycle, and it has roughly the same effect on delay as halving the cycle length. Section 7.2 also shows how reservice can be applied to vehicular left turns as a way of mitigating a change from permitted to protected-only left turns (a treatment described in Section 6.1). The next pair of treatments aims to increase the time that pedestrians can use within a signal cycle. One is Maximizing Walk Interval Length (Section 7.3). Longer Walk intervals both

86 Traffic Signal Control Strategies for Pedestrians and Bicyclists reduce pedestrian delay and make crossings accessible to pedestrians with lower walking speeds. National and local standards specify a minimum Walk interval length (usually 7 s), and all too often, signal timing uses that minimum standard even when a longer Walk interval would fit within the signal cycle without constraining other traffic movements. Pedestrian Clearance Settings for Better Serving Slower Pedestrians (Section 7.4) puts a focus on the final intervals of a pedestrian phase: Flashing Don’t Walk (FDW) and the pedes- trian phase end buffer. It shows how setting these intervals’ lengths to maximize usability can enable a pedestrian phase to have a longer Walk interval and/or support pedestrians with lower walking speeds. The final pair of treatments deals with balancing traffic control’s flexibility, or demand respon- siveness, with its impact on users. Pedestrian Recall versus Actuation (Section 7.5) addresses the long-standing issue of when pedestrian signals should be pushbutton actuated versus on recall (i.e., automatic). Analysis methods used in traditional practice have typically underesti- mated the delay impact to pedestrians and overestimated the impact to vehicles. In many cases, the vehicle delay impact of having pedestrian phases on recall is negligible, indicating pedestrian recall may be warranted even where pedestrian demand is low. Pedestrian Hybrid Beacons (Section  7.6) are an alternative to full traffic signals whose demand responsiveness can enable them to serve both pedestrians and vehicles well, with pedes- trians served only moments after arriving and vehicles stopped only when needed for a crossing pedestrian. 7.1 Short Cycle Length 7.1.1 Basic Description 7.1.1.1 Alternative Names None. 7.1.1.2 Description and Objective The length of a signal cycle is the time from the moment a particular phase ends until that phase is served and ends again. Cycle length may be fixed—as it is with pretimed control and with coordinated-actuated control—or variable, as it is when a signal is running free. For a given set of vehicular and pedestrian demands, there is a minimum cycle length below which capacity will be exceeded due to the capacity loss that occurs when the controller switches phases. A cycle length for a particular situation can be considered short if it is close to that minimum. At a compact intersection with moderate traffic and no turning phases, cycle length can sometimes be as short as 40 s; at more complex intersections with protected turning phases, longer pedestrian crossings, and greater levels of traffic, the minimum cycle length could be 90 s or longer. Shorter cycle lengths reduce delay for pedestrians and bicycles (and usually for transit) because they involve shorter red periods. They reduce crowding on crossing islands and other queuing areas because fewer people cross per cycle. In addition, short cycle lengths have the potential to make roads safer for walking and cycling because they can reduce speeding oppor- tunities, as explained later. 7.1.1.3 Variations Cycle lengths can either be fixed—as they are with coordinated and pretimed control— or variable, as they are with fully actuated control. Fully actuated control, also called running free operation, tends to lead to short cycles because each phase’s green is only extended while traffic is still flowing, given minimum and maximum

Treatments that Reduce Pedestrian and Bicycle Delays 87   green constraints. This allows the intersection to cycle as quickly as possible for the current level of traffic, automatically lengthening cycles when traffic is heavy and shortening them when traffic is light. With coordinated and pretimed control, a common cycle length is applied to all the inter- sections in a corridor or grid for a given coordination period (e.g., a.m. peak). In general, the cycle length is based on the needs of the most demanding intersection during the most demanding 15-minute interval in the coordination period. Therefore, shorter cycle lengths can be achieved by having short coordination zones made up of only one to four intersections as well as short coordination periods. Simple forms of adaptive control constantly monitor traffic volumes and adjust cycle lengths every 10 or 15 minutes. Where a corridor has a long coordination cycle, less demanding intersections may be able to double cycle, meaning they operate with half of the corridor’s prevailing cycle length; this yields the local benefits of a short cycle length without disrupting arterial coordination. 7.1.1.4 Operating Context Short cycle lengths are of interest at every signalized intersection used by pedestrians or bicycles. More specifically: • Where a single, complex intersection demands a substantially larger cycle length than the intersections around it, consider taking it out of coordination—either to run free or to run with its own fixed cycle—so that the intersections around it can operate with shorter cycles. • Where the spacing between two intersections is long enough for queuing with little risk of spillback (around 600 ft, depending on traffic volume and cycle length), coordination zones can be broken to create short coordination zones, with each given the cycle length its most demanding intersection needs. • Where the current signal operation uses a small number of coordination periods in a day (e.g., a.m. peak, midday, p.m. peak, and nighttime), breaking the day into more coordination periods can allow many hours of operation with shorter cycles. • Where a corridor has a long cycle length, simple intersections (i.e., intersections with fewer phases and less cross traffic) can be considered for double cycling. • Where there are one-way grids, such intersections are especially amenable to short cycles, since one-way streets do not require turn phases. For example, downtown Portland, OR, uses cycles of 56 s during off-peak and 60 s during peak periods; the city would use still shorter periods if not for the long clearance time needed by street-running light rail (P. Koonce, personal communication, August 2019). These short cycles contribute to a pedestrian- and bicycle-friendly environment. 7.1.2 Applications and Expected Outcomes 7.1.2.1 National and International Use Many cities have, by policy, a maximum cycle length and an objective to keep cycle lengths as short as possible. In U.S. cities, a common maximum is 120 s; however, while this is far shorter than the 240 s cycles allowed in some suburban jurisdictions, 120 s is not always considered a short cycle. NYC DOT employs 90-second cycles across Manhattan to limit pedestrian delay, with few exceptions (D. Nguyen, personal communication, August 16, 2019). Cambridge, MA, uses 90-second cycles during peak hours and 75-second cycles during off-peak hours (P. Baxter, personal communication, 2019). The policy in Amsterdam, Netherlands, is that cycles should be as short as possible, never exceeding 100 s. One of the city’s core principles for increasing compliance is that signal con- trol should be “credible,” meaning people should not be given a red signal when there is no

88 Traffic Signal Control Strategies for Pedestrians and Bicyclists conflicting traffic. This leads to preference for fully actuated control at most intersections (S. Linders, personal communication, 2018). At compact intersections—even with pedestrian calls—cycle lengths can be as low as 40 s, and they can be even shorter when there are no pedes- trian calls. U.S. practice tends to be strongly in favor of coordination (versus letting signals run free) with long cycles and long coordination zones. Zurich, Switzerland, times its traffic signals using short coordination zones, usually with only two or three intersections per zone. Through traffic gets a green wave through a few intersections followed by a short waiting time (due to the low cycle lengths), which helps compress the platoon and deter speeding (J. Christen & R. Gygli, personal communication, 2005). Coordination periods in the U.S. also tend to be long, with agencies often using far fewer time-of-day plans than their controller software can support. This creates an opportunity to lower cycle lengths by dividing up longer periods when one part of the period has considerably less traffic than another. Fully actuated control is widely used in the U.S., but in many cities, few signals have fully actuated control. By contrast, in Amsterdam and other Dutch cities, most intersections use fully actuated control because it keeps cycles short and is especially suitable for transit signal priority, since it can naturally recover from priority disruptions. With fully actuated control, cycle lengths can be substantially shorter using “snappy” versus “sluggish” settings. Snappy settings include a short unit extension, short minimum green, non- simultaneous gap-out, upstream detectors for gap-out, and lane-by-lane detection and gap-out (Furth et al., 2009). 7.1.2.2 Benefits and Impacts For pedestrians, a shorter signal cycle almost always means less delay. Shorter cycles also improve pedestrian safety, as pedestrian compliance tends to be poor when pedestrians expe- rience long red periods, especially if there are periods with no conflicting traffic (Kothuri et al., 2017). Short cycles usually mean less average delay for vehicles too, as long as the cycle stays long enough to avoid a capacity shortfall. However, with coordinated control, one subset of vehicles—long-distance through traffic—prefers long cycles because they make it possible to create better two-way green waves. As a result, one often sees long green periods in which, for example, the northbound platoon passes through early in the green and the southbound platoon late in the green, with much of the green period being unused in each direction. Meanwhile, cross traffic, left-turning traffic, and transit users (buses cannot stay in the green wave because of stops) have longer delay. Therefore, using shorter cycle lengths to break up long coordination zones and using fully actuated control will mean more delay for long-distance through traffic but less delay for local traffic. In many cases, the net effect is to lower average vehicular delay. Many simulation studies have found that, for the corridor they studied, breaking up coordi- nation zones and either running free or using short-zone coordination with short cycles reduced average vehicular delay as well as pedestrian delay. For example, Kothuri et al. (2017) found that free operation on a corridor in Portland reduced average pedestrian delay from 45 s to 30 s com- pared to coordinated control with fixed cycle lengths, and at the same time it reduced average vehicle delay from 26.5 s to 23 s. Ishaque and Noland (2007) found that on a coordinated arterial, shorter cycle lengths substantially lowered pedestrian delay and yielded small delay-reductions for motorists as well. In some corridors with low pedestrian-demand, designers sometimes face a choice: Either use a long cycle length in which the minor street gets enough time in every cycle to support a

Treatments that Reduce Pedestrian and Bicycle Delays 89   pedestrian phase, or use a shorter cycle in which the minor street is given only the time needed to serve vehicles. In the latter case, when there is a pedestrian call, the minor-street phase runs beyond its scheduled time, and the controller then has to transition over the next one or two cycles to return the intersection to coordination. A study of a Utah corridor by Chowdhury et al. (2019) found that although through traffic had less delay in the long cycle alternative, the short cycle alternative yielded lower vehicular delay overall as well as lower pedestrian delay. Short cycle lengths can also improve the safety of a road by inhibiting speeding. Long cycles tend to have long green periods within which vehicles can speed through several intersections; the faster one drives, the farther one can go before hitting a red light. A study of a Boston cor- ridor compared coordinated control with a 120-second cycle (existing control) to two other control alternatives: fully actuated control and small-zone coordination. In the latter alternative, the corridor was divided into three zones with two or three intersections each. Each zone had its own cycle length; one intersection was in a zone by itself and ran free (Furth et al., 2018). The results, summarized in Exhibit 7-1, showed that both small-zone coordination and run- ning free led to cycle lengths that were more than 30% shorter and yielded large reductions in pedestrian delay. Small-zone coordination also led to a decrease in average vehicular delay. Most importantly, this study measured the number of “speeding opportunities” afforded by the three control alternatives, defined as the number of vehicles arriving at a stop line while the signal is green but with no vehicle less than 5 s ahead of them. Compared to coordinated control, small- zone coordination eliminated more than one-third of speeding opportunities and running free eliminated nearly two-thirds. 7.1.3 Considerations 7.1.3.1 Accessibility Considerations Not applicable for this treatment. 7.1.3.2 Guidance NCHRP Report 812: Signal Timing Manual, 2nd Edition (STM2) advises that, when designing coordination plans, practitioners should consider whether intersections with exceptionally long cycle-length requirements would better operate independently from a group, especially if they are distant enough from neighboring intersections to prevent spillback. The National Association of City Transportation Officials (NACTO) Urban Street Design Guide, PedSafe, and other guides also recommend shorter cycles due to increased efficiency and the potential for increased compliance by all users. 7.1.3.3 Relationships to Relevant Treatments Using pedestrian clearance settings for serving slower pedestrians (see Section  7.4) can shorten cycle length by avoiding periods at the end of a phase that are not needed by either vehicles or pedestrians. Change in average cycle length and average pedestrian delay Change in vehicular delay Change in speeding opportunities Small-zone coordination -33% -13% -37% Fully actuated control (running free) -31% 11% -65% Source: Derived from Furth et al. (2018). Exhibit 7-1. Changes in delay and speeding opportunities compared to coordinated-actuated control.

90 Traffic Signal Control Strategies for Pedestrians and Bicyclists There can be some tension between the desire to keep cycles short and the desire to protect pedestrians from turning conflicts using such techniques as exclusive pedestrian phases (see Section 6.3) and leading pedestrian intervals (LPIs) (see Section 6.5), which can require longer cycle lengths. However, there often are ways to provide the desired protection with little or no cycle-length impact. For example, in most situations, concurrent-protected phasing (see Sec- tion 6.2) requires a far shorter cycle length than exclusive pedestrian phases, while still providing fully protected pedestrian crossings, and delayed turn (see Section 6.6) can provide the same or greater partial protection as an LPI with less cycle-length impact. Changing some aspects of an intersection’s corner geometry can shorten the needed LPI (see Section 6.5), and using overlap- ping pedestrian phases (see Section 6.7) can lessen the cycle-length impact of LPIs. 7.1.3.4 Other Considerations Studies sometimes find that coordination plans with long cycle lengths reduce emissions because they reduce the number of vehicle stops. However, results like this stem from using a short-term analysis framework that assumes fixed travel patterns and ignores human response. Over time, traffic control that reduces travel time for long-distance through traffic will lead to people making longer trips (e.g., by changing their residence or work place) and increasing vehicle-miles traveled, which will drive up emissions by far more than the savings that come from fewer stops. 7.1.4 Implementation Support 7.1.4.1 Equipment Needs and Features Fully actuated control requires detectors on all approaches, while coordinated-actuated con- trol requires detectors on non-coordinated phases only. Pretimed control requires no detectors. 7.1.4.2 Phasing and Timing The minimum necessary cycle length for achieving a target degree of saturation is given by Equation 7-1, derived from the 2016 Highway Capacity Manual: A Guide for Multimodal Mobility Analysis: 1 (7-1)∑ ∑ = −     C L v s X min ci ci ci target where Xtarget = the target degree of saturation (typically in the range 0.85 to 0.95); Cmin = the minimum cycle length necessary to avoid exceeding that target—both sums are over the critical movements only; Lci = the lost time associated with critical phase i; vci = the volume flow rate of critical phase i; and sci = the saturation flow rate of critical phase i. When a pedestrian crossing is critical, its entire phase time should be treated as lost time and its saturation flow rate as infinite. As the cycle-length formula shows, an increase in a critical movement’s lost time generally leads to a far greater increase in necessary cycle length. For example, if the denominator in Equation 7-1 is 0.25, a 4 s increase in critical lost time (as might be caused by an LPI applied to a critical movement) would increase the necessary cycle length by 16 s.

Treatments that Reduce Pedestrian and Bicycle Delays 91   Some controllers have special features that can reduce phase lengths and sometimes reduce cycle lengths. Sobie et al. (2016) describe the “split extension” feature, which allows a coordi- nated phase to terminate early if there are no active vehicles on the approach. This reduces delay for vehicles and for pedestrians waiting to cross. Many forms of adaptive control also adjust cycle length automatically. 7.1.4.3 Signage and Striping Not applicable for this treatment. 7.1.4.4 Geometric Elements The reduction of crossing lengths will allow for shorter cycle lengths when the pedestrian phase is critical. Bibliography American Association of State Highway and Transportation Officials. (2011). A Policy on Geometric Design of Highways and Streets, 7th Edition. Washington, DC. Cesme, B., & Furth, P. G. (2014). Self-Organizing Traffic Signals Using Secondary Extension and Dynamic Coordination. Transportation Research Part C: Emerging Technologies, 48, 1–15. Chowdhury, S.-E.-S., Stevanovic, A., & Mitrovic, N. (2019). Estimating Pedestrian Impact on Coordination of Urban Corridors. Transportation Research Record: Journal of the Transportation Research Board, 2673(7), 265–280. Denney, R. W., Jr., Curtis, E., & Head, L. (2009). Long Green Times and Cycles at Congested Traffic Signals. Transportation Research Record: Journal of the Transportation Research Board, 2128(1), 1–10. Furth, P. G., Cesme, B., & Muller, T. H. J. (2009). Lost Time and Cycle Length for Actuated Traffic Signal. Trans- portation Research Record: Journal of the Transportation Research Board, 2128(1), 152–160. Furth, P. G., Halawani, A. T. M., Li, J., Hu, W., & Cesme, B. (2018). Using Traffic Signal Control to Limit Speeding Opportunities on Bidirectional Urban Arterials. Transportation Research Record: Journal of the Transporta- tion Research Board, 2672(18), 107–116. Gordon, R. L. (2010). NCHRP Synthesis 409: Traffic Signal Retiming Practices in the United States. Transportation Research Board, Washington, DC. Highway Capacity Manual: A Guide for Multimodal Mobility Analysis, 6th Edition. (2016). Transportation Research Board, Washington, DC. Ishaque, M. M., & Noland, R. B. (2007). Trade-Offs between Vehicular and Pedestrian Traffic Using Micro- Simulation Methods. Transport Policy, 14(2), 124–138. Kothuri, S. M., Koonce, P., Monsere, C. M., & Reynolds, T. (2015). Exploring Thresholds for Timing Strategies on a Pedestrian Active Corridor. Kothuri, S. M., Kading, A., Smaglik, E. J., & Sobie, C. (2017). Improving Walkability through Control Strategies at Signalized Intersections. National Association of City Transportation Officials. (2019). Don’t Give Up at the Intersection: Designing All Ages and Abilities Bicycle Crossings. Sobie, C., Smaglik, E., Sharma, A., Kading, A., Kothuri, S., & Koonce, P. (2016). Managing User Delay with a Focus on Pedestrian Operations. Transportation Research Record: Journal of the Transportation Research Board, 2558(1), 20–29. Urbanik, T., Tanaka, A., Lozner, B., Lindstrom, E., Lee, K., Quayle, S., Beaird, S., Tsoi, S., Ryus, P., Gettman, D., Sunkari, S., Balke, K., & Bullock, D. (2015). NCHRP Report 812: Signal Timing Manual, 2nd Edition. Trans- portation Research Board, Washington, DC. 7.2 Reservice 7.2.1 Basic Description 7.2.1.1 Alternative Names None.

92 Trafc Signal Control Strategies for Pedestrians and Bicyclists 7.2.1.2 Description and Objective Reservice refers to serving a trac movement two or more times within a signal cycle. Reservice can be applied to bicycle and pedestrian crossings to reduce their delay as well as to right- and le-turn movements to limit their delay when they are converted to protected-only phasing, with the goal of improving the safety of a bicycle/pedestrian crossing. Exhibit 7-2 shows an intersection in Amsterdam where bicycles and pedestrians are sometimes given two crossing phases in the east–west direction in a cycle. 7.2.1.3 Variations If every phase is served twice per cycle within a coordinated system, this is called double cycling (see Section 7.1 on short cycle lengths). e local intersection has two cycles within one “system cycle.” When a channelized right turn conicts with no trac movement other than its pedestrian/ bicycle crossing, it may be possible to control that pair of movements as its own intersection (i.e., running free). is will allow the right turn and its crossing movements to be served multiple times per cycle. Reservice can also be applied to minor vehicular movements, including le turns and right turns. It can be a way to mitigate longer delays that occur due to making le turns protected-only or applying no turn on red. 7.2.1.4 Operating Context Pedestrian or bicycle reservice might be appropriate: • Where the cycle length is long and the crossing considered for reservice does not require a long phase, such as a bicycle crossing or a short pedestrian crossing; and • At channelized right turns with signal-protected crossings. is context typically involves multistage crossings (see Sections 6.4 and 10.1). Reservice can also be applied to le-turn and right-turn phases to mitigate the large delay increases to those turning movements caused by making them fully protected. is strategy might be appropriate if the corridor has a long signal cycle. (b) Phasing plan * Bicycle/pedestrian phase served only if there is enough time remaining in the cycle before the scheduled start of the N-S phase or when transit priority creates a second tram phase. (a) Intersection layout N Source: Derived from Furth (2019). Exhibit 7-2. Bicycle and pedestrian reservice at the intersection of Sarphatistraat and Weesperplein, Amsterdam. Detail for the north–south phase has been suppressed.

Treatments that Reduce Pedestrian and Bicycle Delays 93   7.2.2 Applications and Expected Outcomes 7.2.2.1 National and International Use Reservice is a well-known technique, although its application is relatively uncommon. It is sometimes used for transit signal priority; for left turns whose turn bay is too short to store the full left-turn demand of a cycle; and for pedestrians and bicycles as well. 7.2.2.2 Benefits and Impacts Reservice can substantially reduce delay for affected movements when a cycle length is long since it approximately halves the maximum red time. The impact to other movements cannot be generalized; however, when reservice is provided by taking time from movements with ample excess capacity, the impact to those other movements can be small. A simulation study of a Boston, MA, intersection with a channelized right turn whose crossing is signalized found that, by serving the channelized right turn and its crossing twice per cycle, average pedestrian delay was reduced by 20 s. With this plan, right turns also get reservice; while they had less total green time, their delay still fell slightly because their red times were shorter. Other traffic movements were unaffected (Furth et al., 2019). Reservice can also be applied to vehicular phases to mitigate delay increases to left turns or right turns due to making turns fully protected (see Section 6.1) or to prohibiting right turn on red (see Section 6.8). When a protected bicycle lane was installed on West 3rd Street in Long Beach, CA, the phasing plan allowed for left turns to be either leading or lagging, as illustrated in Exhibit 7-3. As implemented, the lagging left could be called only if the leading left had been skipped; however, Furth et al. (2014) found that if both leading and lagging lefts were allowed in the same cycle, delay would be substantially reduced for the left turn (meanwhile, delay for pedestrians and bicycles would increase only a little). In preparation for this guidebook, a study of Boston’s Southwest Corridor bicycle path found that converting northbound left turns at two intersections (Heath Street and Cedar Street) from protected-permitted to protected-only phasing would increase average left-turn delay by 46 s using conventional leading left-turn phasing. Meanwhile, with left-turn reservice, left-turn delay would increase by only 14 s. 7.2.3 Considerations 7.2.3.1 Accessibility Considerations Without an accessible pedestrian signal (APS), pedestrians who cannot see the Walk indica- tions will not know about the reservice. The APS may help alert other pedestrians to the oppor- tunity as well. Source: Furth et al. (2014). Exhibit 7-3. Phasing plan in which a protected left turn across a protected bike lane has both a leading and lagging phase. Based on operations along West 3rd Street, Long Beach, CA.

94 Trafc Signal Control Strategies for Pedestrians and Bicyclists 7.2.3.2 Guidance Not applicable for this treatment. 7.2.3.3 Relationships to Relevant Treatments Where crossings of channelized right turns (see Section 6.4) are signalized, reservice for both the crossing and the right turn can substantially reduce delay. 7.2.4 Implementation Support 7.2.4.1 Equipment Needs and Features Not applicable for this treatment. 7.2.4.2 Phasing and Timing Reservice can be implemented at coordinated as well as free running intersections. At coor- dinated intersections with a xed cycle length, reservice can be conditional on having sucient time to t an extra phase, which will depend on when preceding phases terminate. Where channelized right turns are signalized, reservice can be applied to the right turn and its crossing within a coordinated cycle. If a right turn has no other conicts, it is also possible to allow the turn and its crossing to run free as their own intersection. An intersection in Rijswijk, Netherlands, applies reservice using logic that allows a right turn and its crossing to run free for part of a cycle. Where a north–south arterial with a two-way bicycle path on its west side meets on- and o-ramps of the A4 freeway (see Exhibit 7-4), the bicycle path and the southbound right turn conict only with each other and with the north- bound le. During all of the cycle except the northbound le phase, the bicycle path and the southbound right alternate, running freely with phases that can be quite short. As a result, the bike path and right turn are oen served two or three times during a cycle (Furth et al., 2014), resulting in very low delay. 7.2.4.3 Signage and Striping Not applicable for this treatment. 7.2.4.4 Geometric Elements Not applicable for this treatment. (a) (b) Exhibit 7-4. (a) Layout and (b) phasing plan, in which the bicycle crossing and conicting right turn alternate freely for part of the cycle. Junction of Prinses Beatrixlaan with A4 freeway ramps, Rijswijk, Netherlands.

Treatments that Reduce Pedestrian and Bicycle Delays 95   Bibliography Furth, P. G. (2019). Reservice (Two Bike-Ped Phases per Cycle) at an Amsterdam Intersection [Video]. YouTube. https://youtu.be/J-q4HAzIJAg Furth, P. G., Koonce, P. J., Miao, Y., Peng, F., & Littman, M. (2014). Mitigating Right-Turn Conflict with Protected Yet Concurrent Phasing for Cycle Track and Pedestrian Crossings. Transportation Research Record: Journal of the Transportation Research Board, 2438(1), 81–88. Furth, P. G., Wang, Y. D., & Santos, M. A. (2019). Multi-Stage Pedestrian Crossings and Two-Stage Bicycle Turns: Delay Estimation and Signal Timing Techniques for Limiting Pedestrian and Bicycle Delay. Journal of Trans- portation Technologies, 9(4), 489. 7.3 Maximizing Walk Interval Length 7.3.1 Basic Description 7.3.1.1 Alternative Names Rest in Walk. 7.3.1.2 Description and Objective For pedestrian crossings that are concurrent with a parallel vehicular phase, minimum require- ments for the pedestrian phase can often be met with time left over before the vehicular phase ends. This set of treatments aims to add this leftover time to the Walk interval, thereby reducing pedestrian delay, increasing compliance, and making the crossing accessible to slower pedes- trians without significantly constraining the signal cycle. 7.3.1.3 Variations Rest in Walk (for coordinated phases). Rest in Walk is a controller setting that maximizes the length of the Walk interval. For signal controllers in the U.S., this setting can be applied only to coordinated phases, whose start times may vary but whose ending time within a signal cycle is generally fixed. With this setting, the concurrent pedestrian signal dwells in the Walk until the “Walk yield point,” which is the pre-scheduled end of green (“green yield point”) minus the time specified for FDW. This way, if the concurrent vehicular green phase begins earlier than sched- uled, the Walk interval will automatically be lengthened correspondingly. Exhibit 7-5 shows how the pedestrian phase runs with and without the Rest in Walk setting. Maximize the Walk (for phases that are not designated as a coordinated phase). For non-coordinated phases, most controllers do not have a Rest in Walk setting, but the same prin- ciple can be applied: Calculate the longest Walk interval that will not constrain the concurrent vehicular phase, allowing time for the FDW interval. If the concurrent vehicular phase is pretimed, the formula is given in Equation 7-2: ( )( )= − −max , (7-2)W Split FDW t Wveh buffer min where W = the length of the Walk interval based on the Rest in Walk principle; Splitveh = the split duration (i.e., green, yellow, and red clearance) for the concurrent vehicle phase; FDW = the length of the FDW interval; tbuffer = the length of the pedestrian phase end buffer time; and Wmin = the minimum interval allowed by policy (typically 7 s).

96 Traffic Signal Control Strategies for Pedestrians and Bicyclists If the concurrent vehicular phase is actuated, the formula is given in Equation 7-3: ( )( )= − −max , (7-3)W MinSplit FDW t Wveh buffer min where MinSplitveh = the minimum split duration (i.e., minimum green, yellow, and red clearance) for the vehicular phase. For additional discussion on determining the lengths of FDW and the pedestrian phase end buffer, see Section 7.4. Example 1: Suppose a pretimed vehicular phase has a split of 35 s that consists of a green interval of 30 s, a yellow interval of 4 s, and a red clearance interval of 1 s. Suppose also that the time needed for FDW is 9 s, and the phase end buffer will coincide with the yellow and red clearance intervals and, therefore, last 5 s. By policy, the minimum Walk interval is 7 s. Following the logic of Rest in Walk, the length of the Walk interval should be 35 – 9 – 5 = 21 s. Example 2: Suppose the same scenario as Example 1, except the vehicular phase is actuated and has a minimum green interval of 18 s and, therefore, a minimum split of 23 s. In that case, the length of the Walk interval should be 23 – 9 – 5 = 9 s. Adapting minimum green to demand. The minimum green time for actuated vehicular phases is usually short (e.g., many cities use 6 s for turn phases and 10 s for through phases) in order to give the controller freedom as early as possible to end that phase when a gap is detected and switch to the next phase. For the same reason, the Walk interval length for pedestrian phases concurrent with an actuated vehicular phase is usually set at the minimum value (usually 7 s). However, if vehicular demand is such that the phase routinely runs past its minimum green, Source: STM2. Exhibit 7-5. Rest in Walk and Rest in Don’t Walk modes.

Treatments that Reduce Pedestrian and Bicycle Delays 97   then minimum green can be increased with little impact. A suggested rule of thumb is to set the minimum green equal to the 30th percentile green time for the relevant period of the day, then adjust the Walk interval based on Equation 7-3. That way, the minimum green and pedestrian settings will constrain the signal cycle for only 30% of the cycles (and those will be cycles that are well below capacity, with a low risk of creating any significant impact on traffic operations). Example 3: Suppose the same scenario as Example 2, in which the concurrent vehicular phase has a minimum green of 18 s. Suppose traffic demand is high enough during the p.m. peak that the 30th per- centile green interval is 25 s long—in 70% of cycles, the vehicular green runs for at least 25 s. One could then increase the minimum green to 25 s with little impact on signal operations and, following Equa- tion 7-3, the Walk interval could be adjusted to 16 s (7 s longer than before). Example 4: Take the same scenario as Example 3, but suppose traffic demand is high enough during the p.m. peak that the phase runs to maximum green, which is 30 s, in 80% of cycles. That makes the 30th percentile green time equal to 30 s. Increase the minimum green to 30 s (in effect making the phase fixed time) and, following Equation 7-3, adjust the Walk interval to 21 s. Adaptive Walk intervals. This treatment, proposed by Furth and Halawani (2016), is simi- lar to adapting minimum green to demand, except that minimum green is set dynamically on a cycle-by-cycle basis to the 30th percentile green time of the last seven cycles, with the Walk interval adjusted accordingly on a cycle-by-cycle basis using Equation 7-3. That way, the length of the Walk interval is adjusted at all times to the largest value it can have without significantly constraining the signal cycle. 7.3.1.4 Operating Context Maximizing the Walk interval could be considered wherever there are concurrent phases and demand on the concurrent vehicular phase is great enough that the phase’s length is governed by vehicular needs rather than pedestrian needs. The variation that applies depends on how the concurrent vehicular phase is configured: • Coordinated phases: Use the Rest in Walk setting. • Pretimed phases that are not a designated coordinated phase: Set the length of the Walk interval based on Equation 7-2. • Actuated phases, including non-coordinated phases at an intersection with coordinated- actuated control and phases at an intersection with fully actuated control: Set the length of the Walk interval based on Equation 7-3, and consider increasing the minimum green either statically (adapting minimum green to demand) or dynamically (adaptive Walk inter- vals). Note that the strategy of adapting minimum green requires data on the distribution of green time. 7.3.2 Applications and Expected Outcomes 7.3.2.1 National and International Use Rest in Walk is widely used in connection with coordinated phases. Many U.S. cities use Rest in Walk by policy with coordinated phases in order to maximize pedestrians’ allowable walk time and minimize their delay (Kothuri, 2014). For pretimed and non-coordinated phases, many cities make it a policy to maximize their Walk interval lengths using Equation 7-2 or 7-3. For example, in Cambridge, MA, most signals are pretimed, but cycle length and splits vary across the day. Designers calculate and apply the Walk interval for each period using Equation 7-2. Any time a Walk phase displays solid Don’t Walk while the concurrent vehicular phase is still green is considered a signal timing error that must be corrected.

98 Traffic Signal Control Strategies for Pedestrians and Bicyclists Still, many cities do not routinely apply Rest in Walk or maximize the Walk interval for non-coordinated phases. Often, a default Walk interval length (usually 7 s) is used even where a longer Walk would fit within a vehicular phase’s minimum timing. Sometimes signals apply the same Walk interval length for the entire day when, based on Equations 7-2 and 7-3, it could be longer in periods with longer signal cycles. This failure to lengthen Walk intervals even when it would not affect vehicular operations at all is not due to antipathy on the part of designers toward pedestrians; rather, it is a shortcoming of signal timing software and of signal controller design. Commonly used signal timing soft- ware sets Walk intervals to their minimum value by default, rather than applying Equation 7-2 or 7-3; designers often follow the software’s recommendation without checking if longer Walk intervals could be used. And most U.S. signal controllers lack a “Rest in Walk” setting for non- coordinated phases. If they had that setting, maximized Walk intervals would automatically be applied, removing the need for engineers to calculate and set Walk interval lengths for each period of the day. Adaptive Walk intervals require custom programming and have not yet been applied in the U.S. A form of this logic is used in Amsterdam under the name “variable max green” (S. Linders, personal communication, August 1, 2018). 7.3.2.2 Benefits and Impacts Maximizing the length of the Walk interval reduces pedestrian delay, improves pedestrian compliance, and makes crossings accessible to slower pedestrians, with no discernable impact on vehicular traffic. A comparison study found the combination of Rest in Walk and pedestrian recall (see Sec- tion 7.5) increased compliance by bicycles and pedestrians from 9% to 70% for one comparison pair and from 31% to 79% for another comparison pair (Mirabella, 2013). The findings are con- sistent with a second study, which found that pedestrians waiting longer times are more likely to cross illegally (Kothuri et al., 2017). In a simulation study of two intersections, one coordinated-actuated and one fully actuated, adaptive Walk intervals reduced average pedestrian delay by as much as 15 s, with an impact to vehicular traffic of less than 1 s (Furth & Halawani, 2016). Another advantage of adaptive control is that it adjusts to changes in vehicular demand without requiring manual traffic counts. 7.3.3 Considerations 7.3.3.1 Accessibility Considerations The Manual on Uniform Traffic Control Devices (MUTCD, 2009) recommends that where Rest in Walk is applied, the automatic length of the audible Walk interval for APS should be limited to 7 s, while allowing for the audible Walk to restart if the pushbutton is pressed while the visible Walk interval is still timing: MUTCD Section 4E.11 [Standard 5]: The accessible walk indication shall have the same duration as the pedestrian walk signal except when the pedestrian signal is in Rest in Walk. MUTCD Section 4E.11 [Guidance 6]: If the pedestrian signal is in Rest in Walk, the accessible walk indication should be limited to the first 7 seconds of the walk interval. The accessible walk indication should be recalled by a button press during the walk interval provided that the crossing time remaining is greater than the pedestrian change interval. 7.3.3.2 Guidance Not applicable for this treatment.

Treatments that Reduce Pedestrian and Bicycle Delays 99   7.3.3.3 Relationships to Relevant Treatments With most controllers, the Rest in Walk setting automatically applies pedestrian recall (see Section 7.5) to coordinated phases. 7.3.3.4 Other Considerations At intersections with high right-turn volumes and high pedestrian-volumes, it can sometimes be desirable to end the pedestrian phase before the end of the vehicular phase in order to give right-turning traffic a chance to move. Intersections whose coordinated phase is subject to frequent preemption for emergency vehi- cles, railroad vehicles, or other priority vehicles are sometimes exempted from Rest in Walk in order to reduce the likelihood of cutting a pedestrian phase before its clearance has fully timed (Virginia Department of Transportation, 2020). 7.3.4 Implementation Support 7.3.4.1 Equipment Needs and Features Almost all modern traffic signal controllers have a Rest in Walk setting for coordinated phases. The lack of a Rest in Walk setting for non-coordinated phases is a deficiency that should be addressed. Because this setting is lacking, Walk intervals must be calculated and set for each period of the day, even for pretimed phases. Providing a Rest in Walk option to non-coordinated phases could be very easy. The only needed user input would be a minimum Walk interval length; the controller already has the other settings needed to apply Equations 7-2 and 7-3. Adaptive Walk intervals are a form of adaptive control requiring custom logic. However, unlike most forms of adaptive control, it does not require any detection; the only needed input is the duration of recent green intervals. 7.3.4.2 Phasing and Timing An example of phasing and timing for the Rest in Walk feature was provided earlier in Section 7.3.1. 7.3.4.3 Signage and Striping Not applicable for this treatment. 7.3.4.4 Geometric Elements Not applicable for this treatment. Bibliography Furth, P. G., & Halawani, A. T. (2016). Adaptive Walk Intervals. Transportation Research Record: Journal of the Transportation Research Board, 2586(1), 83–89. Kothuri, S. M. (2014). Exploring Pedestrian Responsive Traffic Signal Timing Strategies in Urban Areas. Kothuri, S. M., Kading, A., Smaglik, E. J., & Sobie, C. (2017). Improving Walkability through Control Strategies at Signalized Intersections. Manual on Uniform Traffic Control Devices for Streets and Highways. (2009). FHWA, U.S. DOT. http://mutcd. fhwa.dot.gov/ Mirabella, J. A. (2013). Understanding Pedestrian and Bicyclist Compliance and Safety Impacts of Different Walk Modes at Signalized Intersections for a Livable Community.

100 Traffic Signal Control Strategies for Pedestrians and Bicyclists Urbanik, T., Tanaka, A., Lozner, B., Lindstrom, E., Lee, K., Quayle, S., Beaird, S., Tsoi, S., Ryus, P., Gettman, D., Sunkari, S., Balke, K., & Bullock, D. (2015). NCHRP Report 812: Signal Timing Manual, 2nd Edition. Trans- portation Research Board, Washington, DC. Virginia Department of Transportation. (2020, February 14). Smart Traffic Signal System. https://www.ite.org/ pub/?id=e1bc66ee-2354-d714-513b-cc4f59e84e0d 7.4 Pedestrian Clearance Settings for Better Serving Slower Pedestrians 7.4.1 Basic Description 7.4.1.1 Alternative Names None. 7.4.1.2 Description and Objective The objective of this treatment is to go beyond minimum pedestrian clearance standards and maximize a crossing’s accessibility to slower pedestrians. Additionally, another objective is to maximize efficiency by avoiding unnecessarily lengthening vehicular phases or otherwise inter- fering with the signal cycle. It involves the following settings and policies: • Length of the effective phase end buffer, abbreviated as “effBuffer,” a new concept repre- senting the part of the pedestrian phase end buffer that pedestrians can rely on to finish crossings in comfort; • Pedestrian clearance speeds and the performance measure “lowest pedestrian speed designed”; • Whether a concurrent vehicular yellow can begin while the FDW interval is still timing; and • Whether concurrent vehicular yellow and red clearance intervals count toward needed pedes- trian clearance time. Effective phase end buffer. The MUTCD (2009) states that the pedestrian phase end buffer (i.e., the time from the end of FDW, which is also the start of solid Don’t Walk, until conflicting traffic is released) should be at least 3 s. No maximum is specified; in practice, phase end buffers can last 10 s or longer. From experience, pedestrians know that when FDW ends—which is also when the countdown reaches zero—they have a few more seconds to finish crossing; however, they cannot be expected to know how many more seconds they have at any particular crossing. The effective phase end buffer is defined as the portion of the pedestrian phase end buffer that pedestrians can reasonably rely on to finish their crossing in comfort, and it can therefore count against needed clearance time. This guidebook proposes that the effBuffer be limited to 3 or 4 s—it should be 4 s in cities or regions where pedestrian phase end buffers are almost never shorter than 4 s and 3 s elsewhere. Pedestrian clearance speeds and the performance measure “lowest pedestrian speed designed.” For many years, pedestrian signals were timed using a clearance speed of 4.0 feet per second (ft/s). However, research by Fitzpatrick et al. (2006) found that roughly 20% of young pedestrians and 40% of older pedestrians did not walk this fast (see Exhibit 7-6). Consequently, the MUTCD adopted two lower pedestrian clearance speeds: a primary clearance speed of 3.5 ft/s for pedes- trians who begin crossing up to the last moment of the Walk interval, plus a secondary clearance speed of 3.0 ft/s for slower pedestrians who, aware of their limitation, begin crossing only at the onset of the Walk interval. Primary pedestrian clearance time needed is provided in Equation 7-4: =t D scl,needed p (7-4)

Treatments that Reduce Pedestrian and Bicycle Delays 101   where tcl,needed = primary pedestrian clearance time needed (s); D = crosswalk length, curb to curb (); and sp = primary pedestrian clearance speed (/s). For any specied clearance speed, there will still be pedestrians who walk slower. About 8% of younger adults and 26% of older people walk slower than 3.5 /s, and about 2% of younger adults and 9% of older people walk slower than 3.0 /s. In addition, many children are unable to cross at those clearance speeds. Accommodating slower crossers is an important objective to prevent intersections from becoming barriers to mobility. A performance measure for this aspect of a signalized crossing’s accessibility is the lowest pedestrian speed designed, given by Equation 7-5: = − +   2 , (7-5) , , v max D t D D tpa p eff pb p eff where vpa = lowest pedestrian speed designed (/s); D = crosswalk length, curb to curb (); Dpb = distance from the pushbutton to the departure curb (); and tp,e = eective pedestrian phase length (s). Eective pedestrian phase length is given by Equation 7-6: = + + (7-6),t W FDW effBufferp eff Note: Since publication of this figure, the MUTCD “normal” walking speed has changed to 3.5 ft/s. Source: Fitzpatrick et al. (2006). Exhibit 7-6. Walking speed distribution for young (60 years and under) and old (older than 60 years) crossing pedestrians.

102 Traffic Signal Control Strategies for Pedestrians and Bicyclists where W = length of the Walk interval (s); FDW = length of the FDW interval (s); and effBuffer = length of the effective phase end buffer (s). In Equation 7-5, the first term ensures sufficient time for a pedestrian waiting at the curb and departing within 2 s of the onset of Walk; the second term ensures sufficient time for a pedes- trian waiting at the pushbutton and departing at the onset of Walk. In addition to using vpa as a performance measure, agencies can also specify a target value for vpa and make that target value the secondary clearance speed used to design pedestrian intervals. If design based on the primary speed does not satisfy the secondary clearance requirement, the Walk interval (not the FDW interval) should be lengthened until it is satisfied. The MUTCD (2009) clearance requirement involving a 3.0 ft/s walking speed is roughly the same as specifying a target for vpa of 3.0 ft/s; however, the MUTCD calculation does not limit how much of the pedestrian phase end buffer can count against the needed crossing time. Where the pedestrian pushbutton is closer than 6 ft from the curb, it does not guarantee 2 s for pedestrians to begin their crossing. The MUTCD advises that where the crossing population includes many older adults, young children, or others with low walking speeds, a slower primary clearance speed might be consid- ered. Accordingly, many cities use a primary clearance speed of 3.0 ft/s at crossings near schools and older adult centers, and some apply it in large areas of the city or even citywide. At the same time, cities should consider establishing a slower secondary clearance speed, or, at a minimum, report the accessibility measure vpa to show how the needs of pedestrians who cannot attain the primary clearance speed are identified and designed for as well. Example 1: Suppose a crosswalk is 105 ft long, and primary and secondary clearance speeds are 3.5 and 3.0 ft/s, respectively. Primary clearance time needed (Equation 7-4) is 30 s. Suppose pedestrian timing is 4 s of Walk, 26 s of FDW, and 4 s of effective phase end buffer. If the pushbutton is 5 ft from the curb and Equation 7-5 is applied, then vpa = 3.24 ft/s, which does not meet the secondary clearance target of 3.0 ft/s. If the Walk interval is lengthened to 7 s, vpa falls to 2.97 ft/s, and the secondary clearance target is met. Example 2: Suppose a crosswalk length is 60 ft and there is no pushbutton. Many pedestrians in the area are older adults, so the primary clearance speed is 3.0 ft/s, resulting in a primary clearance need of 20 s. Suppose the pedestrian timing is 7 s of Walk, 16 s of FDW, and 4 s of pedestrian phase end buffer. There is no explicit secondary clearance objective, but citizens would like to know whether people unable to walk 3.0 ft/s will also be accounted for in the timings. If Equation 7-5 is applied, vpa = 2.22 ft/s. Whether a concurrent vehicular yellow can begin while the FDW interval is still timing and whether concurrent vehicular yellow and red clearance intervals can count toward needed pedestrian clearance time. The MUTCD is clear that both of these options are allowed; with them, pedestrian clearance and vehicular timing needs can both be met with the least impact on cycle length and/or the best service to pedestrians for a given cycle length. However, state and local policies sometimes restrict these options. Some agencies make it a policy that vehicular yellow may not begin until FDW ends. This restriction takes away designers’ control over the length of the pedestrian phase end buffer. Yellow times typically range from 3 to 5 s and red clearance times range from 0 to 4 s, which can result in pedestrian phase end buffers of up to 9 s. Because no more than 4 s of the phase end buffer can realistically be counted against pedestrian clearance needs, the best balance of service and efficiency results when pedestrian phase end buffers are uniform, lasting 3 or 4 s. That way, pedestrians will know what to expect when the FDW interval and countdown end. It also avoids the inefficiency of a long phase end buffer that does not improve pedestrian service yet constrains the signal cycle, with negative impacts to pedestrians and others.

Treatments that Reduce Pedestrian and Bicycle Delays 103   Likewise, some agencies do not allow yellow or red clearance times to count toward needed pedestrian clearance time, which forces the FDW to be longer. This restriction is similar to low- ering the pedestrian clearance speed in that it gives pedestrians more time to cross; however, unlike lowering clearance speed, this restriction adds time to the end of the phase that cannot be counted on to serve pedestrians. 7.4.1.3 Variations There are no variations to this treatment. 7.4.1.4 Operating Context This treatment is applicable at every signalized crossing. 7.4.2 Applications and Expected Outcomes 7.4.2.1 National and International Use The MUTCD has established a national primary pedestrian clearance speed of 3.5 ft/s. Many cities have also followed its suggestion of using 3.0 ft/s where crossings have a lot of children or older adults. In the last few years, NYC DOT has classified nearly the entire city as a “senior zone,” and as signals are retimed, it is converting crossings to a primary clearance speed of 3.0 ft/s. At the same time, the city has avoided significant traffic impact by allowing the vehicular yellow time (which in New York typically lasts 3 s) to count toward needed pedestrian clearance. The MUTCD has also established a national secondary clearance speed of 3.0 ft/s. Cities that have chosen to use 3.0 ft/s as the primary clearance speed have not usually specified a lower secondary clearance speed. Many U.S. agencies allow yellow time to count toward pedestrian clearance need, consistent with MUTCD guidance; however, many do not. There are also several agencies that, by policy, do not let FDW overlap with the yellow interval. One reason for this restriction is a common limitation of countdown devices: They often can be configured with a zero point at the onset of yellow or at the end of yellow but not in between (see Section 8.1). The combination of this limi- tation with the requirement that FDW must end when the timer reaches zero effectively forces the FDW interval to end with the start of yellow, unless vehicular red clearance time is at least 3 s or a pedestrian phase overlaps an LPI (see Section 6.7). In Europe, pedestrian phases are structured differently than in the U.S., with a solid green man period, a flashing green man period in which people may still begin to cross but slow pedes- trians are advised not to begin, and then a solid red man clearance time during which conflicting traffic is held (and which continues once traffic is released). These two green intervals together are comparable to our Walk interval, and the pedestrian clearance interval is comparable to our combined FDW interval and phase end buffer. In the Netherlands, primary pedestrian clearance speed (for those who might begin through the last moment of the flashing green man interval) is 1.2 m/s (3.9 ft/s); secondary clearance speed (for those who start only during the solid green man interval) is 1.0 m/s (3.3 ft/s). The minimum length of the solid green man interval is 4 s; with this requirement, pedestrians who depart within the first 2 s of the pedestrian phase can cross at 3.0 ft/s as long as the crossing is no longer than 70 ft, a length that is rarely exceeded. 7.4.2.2 Benefits and Impacts Lower pedestrian clearance speeds give pedestrians more time to cross, making crossings accessible to more people but also making the pedestrian phase longer, which can have negative impacts on vehicle capacity as well as on pedestrians. Most prominently, they sometimes force

104 Traffic Signal Control Strategies for Pedestrians and Bicyclists the signal cycle to be longer, which can increase pedestrian delay (see Section 7.1) and make it less feasible for pedestrian phases to be on recall (see Section 7.5). The two restrictions discussed as part of this treatment—not allowing vehicular yellow to begin until FDW has finished timing and not allowing vehicular yellow and red clearance times to count toward needed pedestrian clearance—have an effect similar to lowering clearance speed. However, they are less efficient because they can result in time at the end of the phase that cannot be counted on to serve pedestrians. To illustrate, consider a crosswalk that is 70 ft long and has a primary pedestrian clearance speed of 3.5 ft/s, resulting in a needed clearance time of 20 s; with a specified Walk window of 7 s, the time needed to serve pedestrians is 27 s. The concurrent vehicular phase has little traffic; its yellow time is 4 s and red clearance is 2 s. The two potential restrictions create four timing alternatives, with impacts shown in Exhibit 7-7 and summarized as follows: A. With both restrictions in place, the vehicular change interval (yellow, red clearance) does not begin until the pedestrian phase has completely cleared, resulting in a phase length of 33 s; this is 6 s more than needed to satisfy pedestrian timing objectives directly. Pedestrians benefit from some of that extra time in that a lower crossing speed is adequate. However, they do not benefit from the last 2 s. B. Letting the yellow begin while FDW is still timing lowers the phase length to 31 s. The unpro- ductive final 2 s of Alternative A are eliminated, with no loss to pedestrian service. C. Letting 4 s of the pedestrian phase end buffer count toward pedestrian clearance needs short- ens FDW by 4 s, and phase length falls to 29 s. The lowest pedestrian speed designed for rises, but it still meets the performance target. Once again, the final 2 s of the phase are of no benefit to pedestrians. D. Relaxing both restrictions allows the phase length to be 27 s, which is what would have nor- mally been calculated as the time needed to serve pedestrians. It yields the same pedestrian performance as Alternative C but with less impact on the signal cycle. This alternative can be seen as first setting the pedestrian signals based on pedestrian needs and then fitting the vehicular signals around that schedule without further constraining the cycle. Alternative C, the second-most efficient, is one that agencies may not have considered before now, as guidance documents have not addressed the idea of allowing the early part, but not the latter part, of the pedestrian phase end buffer to count toward needed pedestrian clearance. This alternative is compatible with countdown devices and can be implemented with nothing more than adjusting the FDW setting. To illustrate impacts when the concurrent vehicular phase is pretimed or coordinated, results for the same example are shown in Exhibit 7-8, with the phase length fixed at 36 s and Alternative A B C D Yellow begins while FDW is still timing? No Yes No Yes Phase end buffer (up to 4 s) counts toward needed clearance? No No Yes Yes Walk interval duration (s) 7 7 7 7 FDW (s) 20 20 16 16 Phase end buffer duration (s) 6 4 6 4 Overall phase duration (s) 33 31 29 27 Lowest pedestrian speed designed, vpa (ft/s) 2.41 2.41 2.80 2.80 Exhibit 7-7. Alternative pedestrian timings for a crossing 70 ft long, with phase timing dominated by pedestrian timing needs.

Treatments that Reduce Pedestrian and Bicycle Delays 105   the cycle length at 90 s. As restrictions are relaxed, the Walk interval increases in length from 10 to 16 s, with corresponding reductions in pedestrian delay. For Alternatives B and D, which allow the yellow to begin while FDW is still timing, the lowest pedestrian speed designed is reduced as well. 7.4.3 Considerations 7.4.3.1 Accessibility Considerations Lower pedestrian design speeds help make crossings accessible to pedestrians with low walking speeds, including older adults and children. Reporting “lowest pedestrian speed accommodated” as a performance measure can help ensure that signal timing is responsive to the needs of slower pedestrians and can reassure citizens that crossings support people who are unable to walk at the primary clearance speed. 7.4.3.2 Guidance Not applicable for this treatment. 7.4.3.3 Relationships to Relevant Treatments Efficient pedestrian clearance settings can enable short cycle lengths (see Section 7.1) and can help maximize Walk interval length (see Section 7.3). 7.4.4 Implementation Support 7.4.4.1 Equipment Needs and Features While most controllers allow FDW to time during the yellow interval, many countdown devices will not extend partway into a yellow phase—they can be set with a zero point at the start of yellow or the end of yellow but not in between (see Section 8.1). Since countdown timers need to end simultaneously with FDW, this limitation can prevent FDW from ending partway through the yellow, which can have impacts on accessibility, pedestrian delay, and cycle length, as discussed earlier. This limitation is something that countdown device manufacturers should be able to correct. 7.4.4.2 Phasing and Timing Not applicable for this treatment. Alternative A B C D Yellow begins while FDW is still timing? No Yes No Yes Phase end buffer (up to 4 s) counts toward needed clearance? No No Yes Yes Walk interval duration (s) 10 12 14 16 FDW (s) 20 20 16 16 Phase end buffer duration (s) 6 4 6 4 Overall phase duration (s) 36 36 36 36 Lowest pedestrian speed designed, vpa (ft/s) 2.19 2.06 2.19 2.06 Exhibit 7-8. Alternative pedestrian timings for a crossing 70 ft long, with fixed phase length.

106 Traffic Signal Control Strategies for Pedestrians and Bicyclists 7.4.4.3 Signage and Striping Not applicable for this treatment. 7.4.4.4 Geometric Elements Shorter crossings, which might be designed using corner bulb-outs, require less clearance time, making signals more efficient and improving accessibility. Bibliography Coffin, A., & Morrall, J. (1995). Walking Speeds of Elderly Pedestrians at Crosswalks. Transportation Research Record: Journal of the Transportation Research Board, 1487, 63. Fitzpatrick, K., Brewer, M. A., & Turner, S. (2006). Another Look at Pedestrian Walking Speed. Transportation Research Record: Journal of the Transportation Research Board, 1982(1), 21–29. Knoblauch, R. L., Pietrucha, M. T., & Nitzburg, M. (1996). Field Studies of Pedestrian Walking Speed and Start-Up Time. Transportation Research Record: Journal of the Transportation Research Board, 1538(1), 27–38. LaPlante, J. N., & Kaeser, T. P. (2004). The Continuing Evolution of Pedestrian Walking Speed Assumptions. ITE Journal, 74(9), 32. Manual on Uniform Traffic Control Devices for Streets and Highways. (2009). FHWA, U.S. DOT. http://mutcd. fhwa.dot.gov/ 7.5 Pedestrian Recall versus Actuation 7.5.1 Basic Description 7.5.1.1 Alternative Names Pedestrian call modes. 7.5.1.2 Description and Objective At signalized intersections with either fully actuated or coordinated-actuated control, pedes- trian phases can be pushbutton actuated or configured on pedestrian recall. With pedestrian actuation, the pedestrian phase is omitted from a cycle unless a pedestrian manually places a call; with pedestrian recall, a call for pedestrian service is placed automatically every cycle. Recall is more convenient and moderately reduces delay because pedestrians arriving during the time scheduled for the Walk interval will be served immediately, whereas with pushbutton actua- tion, the pedestrian phase will have been skipped unless another pedestrian arrived earlier and pushed the button. Actuation is more efficient for signal operations only if pedestrian demand is low (because with high pedestrian-demand the pedestrian phase will usually be called anyway) and if vehicle demand on the concurrent phase is low enough that, absent a pedestrian call, the phase’s green time is not usually long enough to fit a pedestrian phase. 7.5.1.3 Variations Not applicable for this treatment. 7.5.1.4 Operating Context Where signals are pretimed, pedestrian phases should be on recall. Pushbuttons may be pro- vided for accessibility, but they should not be needed to call for service (see Section 8.4). Where signals are coordinated-actuated, Exhibit 7-9 provides suggested guidance, developed for agencies based on the research conducted for this guidebook, to determine whether pedes- trian signals should be actuated. The guidance was developed with the aim of balancing pedes- trian delay with operations efficiency for vehicles. Pedestrian recall should be considered when

Treatments that Reduce Pedestrian and Bicycle Delays 107   pedestrian demand is large enough that there is a call for service in most cycles, as seen on the horizontal axis with total number of pedestrians (for both pedestrian crossings, unless cross streets are on split phase) per cycle. e guidance also considers when the vehicular green on the concurrent vehicle phase is long enough in most cycles that a pedestrian phase would t without unduly extending the cycle length (as seen on the vertical axis). e second condition— regardless of pedestrian demand—almost always applies to the coordinated phase, which is why coordinated phases should usually have pedestrian recall. is condition also applies to non- coordinated phases with high vehicle demand. Finally, where signals are fully actuated, pedestrian actuation is the better choice in most cases because it leads to shorter signal cycles, which reduces delay for pedestrians as well as vehicles. e only case for which pedestrian recall might be appropriate is when a vehicular phase’s average green is long enough, or its minimum green is nearly long enough, to t a pedes- trian phase. 7.5.2 Applications and Expected Outcomes 7.5.2.1 National and International Use Pedestrian recall is always used where signals follow pretimed operation, which is common in areas with high pedestrian-trac, including most downtowns. Outside of downtowns, the most common control type is coordinated-actuated. In many cities, by policy, the coordinated phase (typically the major street) always has pedestrian recall. However, in many cases the pedestrian phase is still actuated, even when the guaranteed phase length for the coordinated phase is more than enough to t a pedestrian phase. Crossings Si de st re et ve hi cu la rg re en as a fr ac ti on of th e gr ee n ne ed ed fo ra p ed es tr ia n ph as e (S SG ) Number of pedestrians per cycle 0. 7 0. 85 1. 0 Pedestrian Recall Pe de st ri an A ct ua tio n 0.4 0.90.0 Exhibit 7-9. Criteria for pedestrian recall versus pedestrian actuation with coordinated-actuated control.

108 Traffic Signal Control Strategies for Pedestrians and Bicyclists associated with a non-coordinated phase are usually actuated, but they may be set to recall where pedestrian demand is high. For example, Boston’s policy is to apply pedestrian recall for a crossing if pedestrians are present for at least 50% of the cycles (City of Boston, 2013). When a side street is set to pedestrian recall, a signal’s operation usually becomes almost pretimed with coordinated-actuated operation; the only demand-actuated phases are the left-turn phases. 7.5.2.2 Benefits and Impacts Where the cycle length is fixed—as with coordinated-actuated control—pedestrian recall reduces pedestrian delay (see Section 3.4, which cites an example in which average pedestrian delay is 10 s greater with actuation than with recall). Lower delay improves pedestrian safety because it tends to improve pedestrian compliance (Otis & Machemehl, 1999; Van Houten et al., 2007). In addition, with pedestrian actuation, pedestrians who are not first to arrive at a corner may not push the button, thinking that it has already been pushed, especially if there is not a prominent call indicator. In that case, if the concurrent vehicular phase receives a green display, there is a risk for pedestrians to cross without the protection of the pedestrian signal, which may leave them partway through their crossing when a conflicting movement is released. Pedestrian recall eliminates this type of conflict and therefore improves safety. For a coordinated phase whose minimum green is long enough to fit a pedestrian phase, there is no delay impact to traffic from having pedestrian recall on the crossing associated with that phase. For non-coordinated phases, pedestrian recall forces the non-coordinated phase to run long enough for the pedestrian phase, which usually constrains the signal cycle and can increase vehicle delay. The impact of pedestrian recall depends on two factors, the first of which is pedes- trian demand. If pedestrian demand is low, recall will constrain most cycles. If demand is high, however, the pedestrian cycle will be called in almost every cycle anyway, in which case recall will have little impact on vehicles. One study conducted using microsimulation recommended that pedestrian signals for side street pedestrians be on recall when there are pedestrian calls in 70% or more of the cycles in a time period and actuated otherwise (Kothuri, 2014). The other relevant factor is traffic volume on the concurrent vehicle phase. If that volume is low, the phase would have a very short green time unless there was a pedestrian call. However, if it is high then the green time would be long (perhaps long enough to fit a pedestrian phase in most cycles), in which case pedestrian recall would have little or no impact. Historically, this factor has rarely been given due attention in discussions about pedestrian recall versus actuation. In production of this guidebook, research was conducted using microsimulation of a cor- ridor in Virginia to measure how the impact of pedestrian recall versus actuation varies with both pedestrian demand and concurrent vehicular demand. As expected, it was found that delay to vehicles from a pedestrian recall setting is greatest when pedestrian volume is low; but once pedestrian volume was so great that there was a pedestrian call in most cycles, that delay impact became negligibly small. Likewise, delay to vehicles due to recall was also greatest when the side street volume was very low; but when side street volume increased such that the average side street green time was at least 75% of the green time needed to fit a pedestrian phase, that delay impact became negligible regardless of the pedestrian demand. This research was the basis for the decision rule given earlier in Exhibit 7-9. 7.5.3 Considerations 7.5.3.1 Accessibility Considerations If a pedestrian phase is on recall then pushbuttons are not needed to call for service, but they can still play a valuable role in making a traffic signal accessible (for more detail, see Section 8.4).

Treatments that Reduce Pedestrian and Bicycle Delays 109   7.5.3.2 Guidance The NACTO Urban Street Design Guide recommends using pretimed signals in urban areas, which results in pedestrian phases being on recall. STM2 indicates that pedestrian recall may be used at locations and/or times with high pedestrian-volumes. Some cities have developed their own guidelines regarding pedestrian recall. For example, Boston recommends pedestrian recall where pedestrians are present for at least 50% of the cycles during peak hours (City of Boston, 2013). 7.5.3.3 Relationships to Relevant Treatments At intersections with fully actuated control, pedestrian actuation helps contribute to short signal cycles (see Section 7.1). 7.5.3.4 Other Considerations Not applicable to this treatment. 7.5.4 Implementation Support 7.5.4.1 Equipment Needs and Features Pedestrian actuation requires pushbuttons, which should be mounted for accessibility and convenience (see Section 8.3). 7.5.4.2 Phasing and Timing Crossings can be set for recall for certain periods of the day and for pedestrian actuation in other periods. With this strategy, call indicators (see Section 8.2) that activate whenever a call has been registered in the current cycle, whether from the pushbutton or automatically, are helpful for informing arriving pedestrians as to whether they need to push the button for service. When pedestrian phases are actuated, signal cycles are usually designed assuming the pedes- trian phase will be served. In operation, when the pedestrian phase is not called, the concurrent vehicular phase ends earlier than scheduled and yields its remaining time to other phases. Where pedestrian demand is low, another option is to design the signal cycle without reserving time for the pedestrian phase, which can allow the overall cycle length to be lower. In operation, when there is a pedestrian call, the concurrent phase runs longer than scheduled, making the next coordinated phase begin late and getting the intersection out of coordination; further control logic is then invoked to recover over the next cycle or two. This way of timing signals can reduce delay for pedestrians as well as vehicles because of the lower cycle length. 7.5.4.3 Signage and Striping Not applicable for this treatment. 7.5.4.4 Geometric Elements Not applicable for this treatment. Bibliography City of Boston. (2013). Boston Complete Streets Design Guidelines. Kothuri, S. M. (2014). Exploring Pedestrian Responsive Traffic Signal Timing Strategies in Urban Areas. Otis, S. C., & Machemehl, R. B. (1999). An Analysis of Pedestrian Signalization in Suburban Areas (No. SWUTC/99/472840-00065-1). Center for Transportation Research, Bureau of Engineering Research, the University of Texas at Austin.

110 Trafc Signal Control Strategies for Pedestrians and Bicyclists Urbanik, T., Tanaka, A., Lozner, B., Lindstrom, E., Lee, K., Quayle, S., Beaird, S., Tsoi, S., Ryus, P., Gettman, D., Sunkari, S., Balke, K., & Bullock, D. (2015). NCHRP Report 812: Signal Timing Manual, 2nd Edition. Trans- portation Research Board, Washington, DC. Van Houten, R., Ellis, R., & Kim, J.-L. (2007). Eects of Various Minimum Green Times on Percentage of Pedes- trians Waiting For Midblock “Walk” Signal. Transportation Research Record: Journal of the Transportation Research Board, 2002(1), 78–83. 7.6 Pedestrian Hybrid Beacons 7.6.1 Basic Description 7.6.1.1 Alternative Names HAWK (high-intensity activated crosswalk) signal. 7.6.1.2 Description and Objective Pedestrian hybrid beacons (PHBs) provide pedestrians with a protected crosswalk without installing a full trac signal. ey include red and yellow aspects but no green aspect, and by default, the beacon is dark. When activated through pedestrian actuation, the yellow light (rst ashing, then solid) warns motorists to stop; then a red light is displayed during which pedes- trians get a Walk signal (see Exhibit 7-10); then there is a clearance period involving a ashing red (to autos) and rst FDW and then solid Don’t Walk for a few seconds (to pedestrians, as the pedestrian phase end buer); and aerward the PHB becomes dark again. Exhibit 7-11 shows a PHB’s display sequence. 7.6.1.3 Variations Not applicable for this treatment. 7.6.1.4 Operating Context PHBs can be applied at unsignalized intersections, midblock crossing locations, and round- about crossings. At intersections, they are used to protect pedestrians crossing the major street (i.e., the street that is not under Stop or Yield control), with beacons facing the major street approaches. PHBs are oen used at intersections whose minor-street demand is too low to war- rant a trac signal, but a safe pedestrian crossing is needed and cannot be achieved with less restrictive measures. PHBs are best suited for: • Multilane crossings (e.g., four lanes or more), particularly those lacking a median refuge island; Source: NACTO. Exhibit 7-10. PHB in Portland providing a protected crossing.

Treatments that Reduce Pedestrian and Bicycle Delays 111   • Crossings of high speed (e.g., 35 mph or more) and high-volume two-lane roadways; • Locations where local-street bicycle routes (also called bicycle boulevards and neighborhood greenways) cross arterials. PHBs have also been favored at these locations over regular trac signals because of neighborhood concerns that regular trac signals might increase trac on the minor street—typically, bicycles need to use pushbuttons for actuation; • Bus stops that lack a safe crossing; • Crossing locations deemed high-risk areas (e.g., schools, shopping centers); and • Crossings with a large number of vulnerable users (e.g., children, elderly, or disabled). PHBs can be coordinated with adjacent signalized intersections, or they can operate in isola- tion. With isolated operation, pedestrians get almost instantaneous service (except for the need to guarantee a minimum “dark” time for mainline between two successive activations), resulting in near-zero delay for pedestrians. Where PHBs are coordinated with adjacent signalized inter- sections, there is a xed window for pedestrians each cycle to receive a Walk indication in order to maintain coordination for vehicles. is typically results in longer pedestrian delay. 7.6.2 Applications and Expected Outcomes 7.6.2.1 National and International Use PHBs were rst installed in the 1990s in Arizona as an adaptation of the British “pelican” pedestrian signal. To date, PHBs can be found across America, with the vast majority at inter- sections. According to a 2019 study conducted by DeLorenzo et al., 41 states have installed PHB devices, seven additional states allow installation of PHBs but have none installed, and one state—Pennsylvania—prohibits PHB installation. (West Virginia did not provide a response in the study.) (DeLorenzo et al., 2019). e lack of PHB implementation for some states is due to concerns with motor vehicle codes that require drivers to stop at dark signals. e same study found that 39 states currently have laws on dark signals, and the other ve states require approaching vehicles to proceed with caution. FHWA is careful to call PHBs “beacons” and not “trac signals.” Some PHBs involve a two-stage crossing, with a separate PHB controlling either side of a divided roadway. 7.6.2.2 Benets and Impacts Studies of PHBs generally show decreased crash rates, both for total crashes and pedestrian- related crashes, aer an unsignalized crossing was converted to a PHB. A before-and-aer study Source: MUTCD (2009), Figure 4F-3. Exhibit 7-11. Phase sequences for PHBs.

112 Traffic Signal Control Strategies for Pedestrians and Bicyclists of 21 PHBs in Tucson, AZ, found that total crash rate and pedestrian crash rate were reduced by 35% and 86%, respectively. During the same before-and-after analysis, a control group of 36 sig- nalized intersections saw a 16% reduction in both total crashes and pedestrian-related crashes, and a control group of 102 unsignalized intersections saw a 9% reduction in total crash rate and a 143% increase in pedestrian crash rate, indicating the effectiveness of PHBs in improving pedestrian safety, especially compared to unsignalized crossings (Fitzpatrick & Park, 2010). A study of PHBs with advanced yield or stop markings or signs from 27 sites resulted in a crash modification factor of 0.244 for pedestrian crashes and 0.82 for all crashes compared to unsignal- ized crossings (Zegeer et al., 2017). Studies have generally found that the motorist yielding rate at PHBs is only a bit lower than yielding rates at full traffic signals. A study of PHBs at midblock crossings in Lawrence, Kansas, showed a driver-yielding rate ranging from 90% to 95% at PHBs compared to 99% at signalized, midblock crossings (Godavarthy, 2010). A study of 20 PHBs in Arizona and Texas, most of them at intersections, found a yielding rate of 96% (Fitzpatrick & Pratt, 2016). The same research of PHBs in Arizona and Texas by Fitzpatrick and Pratt indicated that some drivers may not understand all the signal phases. In particular, 5% of drivers stopped and remained stopped during the flashing red phase, not realizing that they may advance after stop- ping during the flashing red. The study found that only 7% of pedestrians crossed the roadway during the dark indication (Fitzpatrick & Pratt, 2016). With respect to the impact on drivers, the analysis of two PHB sites found significantly less unnecessary delay compared to a signalized midblock crossing (Godavarthy, 2010). “Unneces- sary delay” was defined as time during which the driver was required to stop at the crossing but no pedestrian was present. PHBs with APS installed at a roundabout to increase accessibility and safety for pedestrians with vision disabilities resulted in an intervention rate (i.e., frequency of someone who is visu- ally impaired stepping out and needing to be stopped to prevent collision) of 0% compared to interventions of 2.4% to 2.8% on two-lane roundabouts with similar crossings without PHB. The intervention rate was 0.8% to 1.4% on single-lane roundabouts without PHB (Schroeder et al., 2011). The time between pedestrian actuation and the Walk interval given to pedestrians should be carefully considered and minimized. Excessive delay can result in noncompliance by pedestrians, which may result in noncompliance by drivers who arrive at a solid red indication without pedes- trians preparing to cross. PHBs that are in isolated operation typically result in very low pedes- trian delay because the Walk interval is provided almost instantaneously. Even on coordinated arterials, if the distance to the nearest coordinated signal is large enough to prevent queue spill- backs, PHBs can be configured in isolated mode to reduce pedestrian delay without causing undue delay for vehicles. Where a PHB is operating in coordinated mode on a coordinated arte- rial with a long cycle length, double cycling can reduce pedestrian delay and increase compliance (see Section 7.2) and will likely have little delay impact on the arterial. 7.6.3 Considerations 7.6.3.1 Accessibility Considerations Pedestrians who are visually impaired or have low vision tend to initiate crossing when they hear the perpendicular traffic stop and the traffic parallel to the crossing start to move. Since traffic parallel to a PHB crosswalk is not signalized, pedestrians will not hear the expected sound of the parallel traffic. Therefore, it is important to include APS that audibly communicate the pedestrian signal indications.

Treatments that Reduce Pedestrian and Bicycle Delays 113   The MUTCD’s recommendations on pedestrian pushbutton location should be followed. Detectable warning surfaces and ADA-accessible curb ramps that apply at signalized crossings apply equally with PHBs. PHBs are recommended by NCHRP Research Report 834: Crossing Solutions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities: A Guidebook as one solu- tion to provide access for pedestrians with vision disabilities at multilane roundabout crosswalks (Schroeder et al., 2017). 7.6.3.2 Guidance MUTCD, Chapter 4F (2009) provides guidance for the application, design, and operation of PHBs. The guidelines provided are based on pedestrian hourly volumes, vehicle hourly volumes, vehicle speed, and the length of the crosswalk. 7.6.3.3 Relationships to Relevant Treatments Pedestrian countdown (see Section 8.1), call indicators (see Section 8.2), and independently mounted pushbuttons (see Section 8.3) are helpful at PHBs, just as at signalized crossings. 7.6.3.4 Other Considerations Not applicable for this treatment. 7.6.4 Implementation Support 7.6.4.1 Equipment Needs and Features PHB installations include a face with two circular red signal indications located side by side, with a circular yellow indication centered below. A second beacon face is either mounted over- head or on the other side of the street. A pair of beacons should be installed for each approach of the major street. A pedestrian signal head and detection equipment are also needed on both ends of the cross- walk. The detection equipment can be independently mounted or mounted on the pole used to support the beacons, if appropriately located for pedestrian access. If a median is present and two-stage crossing is implemented, an additional pair of pedestrian signal heads and detection equipment are necessary for the median. 7.6.4.2 Phasing and Timing The display sequence for PHBs is shown in Exhibit 7-11. During Interval 5, also known as the pedestrian clearance interval, the pedestrian display shows FDW for most of the interval, then Don’t Walk for the final 3 s as a pedestrian phase end buffer. When the beacon is inactive, the pedestrian indication rests in solid Don’t Walk. The duration of the solid yellow indication can be calculated using the typical procedure for a vehicular change interval. The pedestrian Walk, FDW, and pedestrian phase end buffer should be timed as detailed in Sections 4.2 and 7.4. 7.6.4.3 Signage and Striping The MUTCD (2009) requires that a crosswalk and stop lines be installed in conjunction with the PHB. Additionally, a Crosswalk Stop on Red sign (R10-23) must also be mounted. Some PHB installations include an educational plaque mounted near the pedestrian detection. Some PHBs across the country are installed with advanced yield or stop markings and signs. The use of advanced markings increases the distance between the crosswalk and yielding vehicles.

114 Traffic Signal Control Strategies for Pedestrians and Bicyclists This helps prevent vehicles in one lane from screening pedestrians from drivers in other lanes during the flashing red interval, in which vehicles are allowed to advance. 7.6.4.4 Geometric Elements Not applicable for this treatment. Bibliography DeLorenzo, S. S., Jiang, M., & Attalla, M. (2019). Current Policies throughout the Nation for Pedestrian Hybrid Beacon (PHB) Installation. Illinois Center for Transportation/Illinois Department of Transportation. Fitzpatrick, K., & Park, E. S. (2010). Safety Effectiveness of the HAWK Pedestrian Crossing Treatment (No. FHWA- HRT-10-045). FHWA. Fitzpatrick, K., & Pratt, M. P. (2016). Road User Behaviors at Pedestrian Hybrid Beacons. Transportation Research Record: Journal of the Transportation Research Board, 2586(1), 9–16. Godavarthy, R. P. (2010). Effectiveness of a Pedestrian Hybrid Beacon at Mid-Block Pedestrian Crossings in Decreasing Unnecessary Delay to Drivers and a Comparison to Other Systems [Doctoral dissertation, Kansas State University]. Manual on Uniform Traffic Control Devices for Streets and Highways. (2009). FHWA, U.S. DOT. http://mutcd. fhwa.dot.gov/ Schroeder, B., Hughes, R., Rouphail, N., Cunningham, C., Salamati, K., Long, R., Guth, D., Emerson, R. W., Kim, D., Barlow, J., Bentzen, B. L., Rodegerdts, L., & Myers, E. (2011). NCHRP Report 674: Crossing Solu- tions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities. Transportation Research Board, Washington, DC. Schroeder, B., Rodegerdts, L., Jenior, P., Myers, E., Cunningham, C., Salamati, K., Searcy, S., O’Brien, S., Barlow, J., & Bentzen, B. L. (2017). NCHRP Research Report 834: Crossing Solutions at Roundabouts and Chan- nelized Turn Lanes for Pedestrians with Vision Disabilities: A Guidebook. Transportation Research Board, Washington, DC. Zegeer, C., Lyon, C., Srinivasan, R., Persaud, B., Lan, B., Smith, S., Carter, D., Thirsk, N. J., Zegeer, J., Ferguson, E., Van Houten, R., & Sundstrom, C. (2017). Development of Crash Modification Factors for Uncontrolled Pedestrian Crossing Treatments. Transportation Research Record: Journal of the Transportation Research Board, 2636(1), 1–8.

Next: Chapter 8 - Treatments Offering Added Information and Convenience »
Traffic Signal Control Strategies for Pedestrians and Bicyclists Get This Book
×
 Traffic Signal Control Strategies for Pedestrians and Bicyclists
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

In the United States, traffic signal timing is traditionally developed to minimize motor vehicle delay at signalized intersections, with minimal attention paid to the needs of pedestrians and bicyclists. The unintended consequence is often diminished safety and mobility for pedestrians and bicyclists.

The TRB National Cooperative Highway Research Program's NCHRP Research Report 969: Traffic Signal Control Strategies for Pedestrians and Bicyclists is a guidebook that provides tools, performance measures, and policy information to help agencies design and operate signalized intersections in a way that improves safety and service for pedestrians and bicyclists while still meeting the needs of motorized road users.

Supplemental to the report are presentations of preliminary findings, strategies, and summary overview.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!