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.
Chapter 12. Oversaturated Condi ons CHAPTER 12 OVERSATURATED CONDITIONS CONTENTS 12.1 SYMPTOMS OF OVERSATURATION ............................................................................ 12-2 12.1.1 Overlow Queue .................................................................................................................... 12-2 12.1.2 Approach Spillback .............................................................................................................. 12-3 12.1.3 Storage Bay Spillback .......................................................................................................... 12-3 12.1.4 Storage Bay Blocking .......................................................................................................... 12-3 12.1.5 Starvation ................................................................................................................................ 12-4 12.1.6 Cross Intersection Blocking ............................................................................................. 12-4 12.2 OPERATIONAL OBJECTIVES FOR OVERSATURATED CONDITIONS ................. 12-5 12.2.1 Maximizing Intersection Throughput .......................................................................... 12-5 12.2.2 Queue Management ............................................................................................................. 12-5 12.3 MITIGATION STRATEGIES FOR OVERSATURATED CONDITIONS .................... 12-5 12.3.1 Mitigation Strategies for Individual Intersections .................................................. 12-6 12.3.2 Mitigation Strategies for Arterials .............................................................................. 12-11 12.3.3 Mitigation Strategies for Networks ............................................................................ 12-19 12.3.4 Adaptive Control ................................................................................................................ 12-21 12.4 REFERENCES .................................................................................................................... 12-21 Signal Timing Manual, Second Edion
Chapter 12. Oversaturated Condi ons LIST OF EXHIBITS Exhibit 12-1 Oversaturation Performance Periods................................................................ 12-1 Exhibit 12-2 Oversaturation Symptoms ..................................................................................... 12-2 Exhibit 12-3 Overflow Queue .......................................................................................................... 12-2 Exhibit 12-4 Approach Spillback ................................................................................................... 12-3 Exhibit 12-5 Storage Bay Spillback ............................................................................................... 12-3 Exhibit 12-6 Storage Bay Blocking ................................................................................................ 12-4 Exhibit 12-7 Starvation ...................................................................................................................... 12-4 Exhibit 12-8 Cross Intersection Blocking ................................................................................... 12-4 Exhibit 12-9 Oversaturation Mitigation Strategies ................................................................ 12-6 Exhibit 12-10 Example of Cycle Length as a Function of Split Ratio and Link Distance ......................................................................................................................... 12-8 Exhibit 12-11 Ring-and-Barrier Diagram for Re-Service of Left-Turn Phase ................ 12-9 Exhibit 12-12 Merging of Two Heavy Upstream Movements ............................................ 12-10 Exhibit 12-13 Ring-and-Barrier Diagram for Double Cycling ........................................... 12-10 Exhibit 12-14 Phase Truncation Example ................................................................................. 12-11 Exhibit 12-15 Conditions for Leading and Lagging Left-Turn Phases ........................... 12-12 Exhibit 12-16 Example Green Flush Sequence ........................................................................ 12-13 Exhibit 12-17 Average Travel Time Using Different Oversaturation Mitigation Strategies .................................................................................................................... 12-14 Exhibit 12-18 Example of Simultaneous Offsets ..................................................................... 12-15 Exhibit 12-19 Example of Negative Offsets ............................................................................... 12-16 Exhibit 12-20 Example of Offset to Prevent Queue Spillback ............................................ 12-17 Exhibit 12-21 Example of Offset to Prevent Starvation ....................................................... 12-18 Exhibit 12-22 Offset Values to Avoid Spillback and Starvation for Intersections Spaced at 800 Feet .................................................................................................. 12-19 Exhibit 12-23 Example of Gating Operation ............................................................................. 12-20 Signal Timing Manual, Second Edion
Chapter 12. Oversaturated Condi ons LIST OF EXHIBITS Exhibit 12-1 Oversaturation Performance Periods................................................................ 12-1 Exhibit 12-2 Oversaturation Symptoms ..................................................................................... 12-2 Exhibit 12-3 Overflow Queue .......................................................................................................... 12-2 Exhibit 12-4 Approach Spillback ................................................................................................... 12-3 Exhibit 12-5 Storage Bay Spillback ............................................................................................... 12-3 Exhibit 12-6 Storage Bay Blocking ................................................................................................ 12-4 Exhibit 12-7 Starvation ...................................................................................................................... 12-4 Exhibit 12-8 Cross Intersection Blocking ................................................................................... 12-4 Exhibit 12-9 Oversaturation Mitigation Strategies ................................................................ 12-6 Exhibit 12-10 Example of Cycle Length as a Function of Split Ratio and Link Distance ......................................................................................................................... 12-8 Exhibit 12-11 Ring-and-Barrier Diagram for Re-Service of Left-Turn Phase ................ 12-9 Exhibit 12-12 Merging of Two Heavy Upstream Movements ............................................ 12-10 Exhibit 12-13 Ring-and-Barrier Diagram for Double Cycling ........................................... 12-10 Exhibit 12-14 Phase Truncation Example ................................................................................. 12-11 Exhibit 12-15 Conditions for Leading and Lagging Left-Turn Phases ........................... 12-12 Exhibit 12-16 Example Green Flush Sequence ........................................................................ 12-13 Exhibit 12-17 Average Travel Time Using Different Oversaturation Mitigation Strategies .................................................................................................................... 12-14 Exhibit 12-18 Example of Simultaneous Offsets ..................................................................... 12-15 Exhibit 12-19 Example of Negative Offsets ............................................................................... 12-16 Exhibit 12-20 Example of Offset to Prevent Queue Spillback ............................................ 12-17 Exhibit 12-21 Example of Offset to Prevent Starvation ....................................................... 12-18 Exhibit 12-22 Offset Values to Avoid Spillback and Starvation for Intersections Spaced at 800 Feet .................................................................................................. 12-19 Exhibit 12-23 Example of Gating Operation ............................................................................. 12-20 Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-1 CHAPTER 12. OVERSATURATED CONDITIONS Oversaturated conditions occur when demand exceeds capacity, which is frequently observable with heavy congestion and the presence of long queues. Under such conditions, operational objectives are commonly maximizing throughput and managing queues (1), but practitioners may Âind it difÂicult to develop effective trafÂic signal timing to accommodate those operational strategies. Although the issue of oversaturation is often not resolvable solely through new signal timing, there are several mitigation strategies that can be applied to improve overall system performance and increase short-term capacity. When oversaturated conditions are present, performance is typically evaluated during three periods: the loading period, the oversaturated period, and the recovery period. These performance periods (see Exhibit 12-1) are assigned based on trafÂic load and, depending on the operating environment, may last for several minutes to several hours. When oversaturated conditions persist for an extended period of time, it may be beneÂicial to apply different timing strategies during the three performance periods. Characterizing the situation is the Âirst step when selecting a mitigation strategy. This requires the practitioner to deÂine the type of oversaturation, using the following characteristics: ⢠Number of affected intersections, ⢠Number of affected directions of travel, ⢠Duration of oversaturation, ⢠Degree of change over time, ⢠Frequency of oversaturation, ⢠Causes of oversaturation, and ⢠SpeciÂic symptoms of oversaturation. Exhibit 12-1 Oversatura on Performance Periods Signal Timing Manual, Second Edi on
12-2 Chapter 12. Oversaturated Condi ons The following sections will help practitioners de ine the type of oversaturation that is present, the potential operational objectives speci ic to oversaturation, and the operational strategies for accomplishing those objectives. 12.1 SYMPTOMS OF OVERSATURATION Exhibit 12-2 provides a summary of many of the common symptoms that indicate oversaturated conditions. One of the reasons that oversaturation is such a challenging problem for traf ic signal control is that these symptoms often occur in combination. Identifying symptoms (and the degree to which they are present) is important when selecting appropriate operational objectives and mitigation strategies. Additional information about each oversaturation symptom is provided throughout this section. Oversatura on Symptom Descrip on Overflow Queue Part of queue not served during a single cycle Approach Spillback Upstream queue physically blocks downstream vehicles Storage Bay Spillback Turning queue fills storage bay and physically blocks through movement Storage Bay Blocking Through queue physically blocks turning movement from storage bay Starva on Traffic demand is restricted from using full downstream roadway capacity Cross Intersec on Blocking Queue extends into an intersec on and blocks progression of crossing vehicles 12.1.1 Overflow Queue An over low queue is one of the most easily observed indications of oversaturated conditions. It is de ined as the part of a queue that is not processed during the green interval and that must be served by subsequent cycles (as shown in Exhibit 12-3). Exhibit 12-2 Oversatura on Symptoms Exhibit 12-3 Overflow Queue Signal Timing Manual, Second Edi on
12-2 Chapter 12. Oversaturated Condi ons The following sections will help practitioners de ine the type of oversaturation that is present, the potential operational objectives speci ic to oversaturation, and the operational strategies for accomplishing those objectives. 12.1 SYMPTOMS OF OVERSATURATION Exhibit 12-2 provides a summary of many of the common symptoms that indicate oversaturated conditions. One of the reasons that oversaturation is such a challenging problem for traf ic signal control is that these symptoms often occur in combination. Identifying symptoms (and the degree to which they are present) is important when selecting appropriate operational objectives and mitigation strategies. Additional information about each oversaturation symptom is provided throughout this section. Oversatura on Symptom Descrip on Overflow Queue Part of queue not served during a single cycle Approach Spillback Upstream queue physically blocks downstream vehicles Storage Bay Spillback Turning queue fills storage bay and physically blocks through movement Storage Bay Blocking Through queue physically blocks turning movement from storage bay Starva on Traffic demand is restricted from using full downstream roadway capacity Cross Intersec on Blocking Queue extends into an intersec on and blocks progression of crossing vehicles 12.1.1 Overflow Queue An over low queue is one of the most easily observed indications of oversaturated conditions. It is de ined as the part of a queue that is not processed during the green interval and that must be served by subsequent cycles (as shown in Exhibit 12-3). Exhibit 12-2 Oversatura on Symptoms Exhibit 12-3 Overflow Queue Signal Timing Manual, Second Edi on Chapter 12. Oversaturated Condi ons 12-3 12.1.2 Approach Spillback Approach spillback occurs when the queue from a downstream intersection occupies all available space on a link and prevents upstream vehicles from entering the downstream link on green (see Exhibit 12-4). Some literature has deÂined this condition as âde facto redâ for the upstream movement because no progression is possible. The oversaturation cause is Intersection B, and simple timing adjustments at Intersection A (without understanding the problem at Intersection B) are likely to be ineffective. 12.1.3 Storage Bay Spillback Storage bay spillback (shown in Exhibit 12-5) occurs when turning trafÂic uses up the entire space of the storage bay and blocks through trafÂic. The blocked through movement may then experience starvation (explained in Section 12.1.5). Storage bay spillback is commonly caused by inadequate green time for the turning phase or ineffective phase sequencing. 12.1.4 Storage Bay Blocking Storage bay blocking (illustrated in Exhibit 12-6) is the converse of storage bay spillback. Storage bay blocking occurs when queues for a through movement extend beyond the opening of a storage bay. In this situation, the turning movement experiences starvation (explained in Section 12.1.5) because left- or right-turning vehicles are blocked. If the storage bay is unoccupied because through vehicles have blocked all turning trafÂic, the turning movement may be skipped entirely. This condition is commonly caused by inadequate green time for the through movement or by ineffective phase sequencing. Exhibit 12-4 Approach Spillback Exhibit 12-5 Storage Bay Spillback Signal Timing Manual, Second Edi on
12-4 Chapter 12. Oversaturated Condi ons 12.1.5 Starvaon Starvation occurs during the green interval when the use of full roadway capacity is restricted by the effects of spillback, storage bay blocking, or perhaps because the upstream signal is red. In this situation, one of these limiting factors starves the downstream roadway of trafic demand (as illustrated in Exhibit 12-7). This is commonly a result of improper offsets or phase sequence. 12.1.6 Cross Intersecon Blocking Cross intersection blocking (illustrated in Exhibit 12-8) occurs when queues extend into an intersection and block the progression of crossing vehicles. While most jurisdictions have âdonât block the boxâ laws or policies, these situations are not uncommon in grids and networks with short link lengths. Carefully controlled green time settings and signal offsets are necessary when mitigating this scenario. Exhibit 12-6 Storage Bay Blocking Exhibit 12-7 Starva on Exhibit 12-8 Cross Intersec on Blocking Signal Timing Manual, Second Edion
12-4 Chapter 12. Oversaturated Condi ons 12.1.5 Starvaon Starvation occurs during the green interval when the use of full roadway capacity is restricted by the effects of spillback, storage bay blocking, or perhaps because the upstream signal is red. In this situation, one of these limiting factors starves the downstream roadway of trafic demand (as illustrated in Exhibit 12-7). This is commonly a result of improper offsets or phase sequence. 12.1.6 Cross Intersecon Blocking Cross intersection blocking (illustrated in Exhibit 12-8) occurs when queues extend into an intersection and block the progression of crossing vehicles. While most jurisdictions have âdonât block the boxâ laws or policies, these situations are not uncommon in grids and networks with short link lengths. Carefully controlled green time settings and signal offsets are necessary when mitigating this scenario. Exhibit 12-6 Storage Bay Blocking Exhibit 12-7 Starva on Exhibit 12-8 Cross Intersec on Blocking Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-5 12.2 OPERATIONAL OBJECTIVES FOR OVERSATURATED CONDITIONS When conditions are oversaturated, the operational objective is usually to maximize throughput or to manage queues. Minimizing user delay is not typically a viable objective in oversaturated conditions, as it is no longer possible to avoid phase failures. Maximizing throughput and minimizing queues, however, keep as much of the system operational as possible. 12.2.1 Maximizing IntersecÂon Throughput Intersection throughput is deÂined as the number of vehicles that can be served by an intersection over a speciÂied period of time when continuous demand exists. It is typically assessed using two measurements: ⢠The total number of vehicles âinputâ to a system of intersections and ⢠The total number of vehicles âoutputâ by a system of intersections. When running efÂiciently, a signal network will experience an input rate and output rate that are nearly equal. In oversaturated conditions, however, the output rate is less than the input rate, and overÂlow queues begin to build at various points within the system. Maximizing throughput manages the formation and growth of queues, which prevents spillback and resultant congestion. Situations do arise in which pervasive queues will continue to grow until demand diminishes, and further revisions to signal timing are futile. However, the practitioner should consider the following strategies (described in detail throughout Section 12.3) during the performance periods: ⢠Loading PeriodâMaximizing both input and output. ⢠Oversaturated PeriodâUsing available physical capacity to maximize input. ⢠Recovery PeriodâDissipating queues to maximize output. 12.2.2 Queue Management When maximizing throughput fails to relieve growing residual queues, many practitioners change their objective to managing queues. Effective queue management is accomplished through the use of signal timing strategies that allow queues to form in locations where they are least likely to have a detrimental effect on the overall operation of the corridor or network. SpeciÂic strategies for queue management are outlined in Section 12.3. 12.3 MITIGATION STRATEGIES FOR OVERSATURATED CONDITIONS This section describes intersection, arterial, and network mitigation strategies that can be used to relieve oversaturated conditions. Exhibit 12-9 provides a summary of the various mitigation strategies and whether their primary objective is to maximize intersection throughput or manage queues. Depending on the system and oversaturated conditions, these strategies may be used in combination or sequence. Signal Timing Manual, Second EdiÂon
12-6 Chapter 12. Oversaturated Condi ons Operaonal Objecve Type of System Migaon Strategy Definion Maximize Intersec on Throughput Individual Intersec on(s) Split Realloca on Realloca ng split me from under-saturated phases to oversaturated phases. Cycle Length Increase Adding split me to oversaturated phases without reducing split me for minor movements, effec vely increasing the cycle length. Opera on of Closely Spaced Intersec ons on One Controller Opera ng two intersec ons using one controller, allowing close coordina on. Arterial Phase Sequence Modificaon for Le Turns1 Changing the phase sequence to lead or lag le turns, increasing the bandwidth for progressing platoons. Preempon Flushing (âGreen Flushâ) Progressing oversaturated movements by placing calls from downstream intersecons to upstream intersecons. Alternave Timing Plan Flushing Using a ming plan to increase the cycle length and give the majority of the green me to the oversaturated phases. Manage Queues Individual Intersecon(s) Green Extension Adding green me to a phase when a detector exceeds a defined threshold of occupancy. Phase Re-Service Serving an oversaturated phase twice during the same cycle. Phase Truncaon Terminaon of a green interval (despite demand) when there is minimal to no flow over a detector. Arterial Simultaneous Offsets SeÂng offset to zero between intersecons, so that queues move simultaneously. Negave Offsets Starng the green interval earlier at downstream intersecons to allow the downstream queue to dissipate before upstream vehicles arrive. Offsets to Prevent Queue Spillback Choosing offsets that allow the downstream queue to clear before the upstream vehicles reach the end of the downstream queue. Offsets to Prevent Starvaon Choosing offsets that ensure the first released vehicle at the upstream intersecon joins the discharging queue as it begins to move. Network Metering (Gang) Impeding traffic at appropriate upstream points to prevent traffic flow from reaching crical levels at downstream intersecons. Maximize Intersecon Throughput and Manage Queues All Adapve Control Applying detecon data and adapve signal control algorithms to adjust signal ming. Combinaon of Migaon Strategies Combining migaon strategies and/or using them in sequence throughout the oversaturated period. 1 This strategy can also be applied at the individual intersecon level, but is primarily used for arterials. 12.3.1 Migaon Strategies for Individual Intersecons Some strategies are best utilized at individual intersections when mitigating oversaturated conditions. 12.3.1.1 Split Reallocaon When oversaturated conditions exist, it is often necessary to reallocate splits to provide a larger portion of the cycle to oversaturated phases and shorter splits to Exhibit 12-9 Oversaturaon Migaon Strategies Signal Timing Manual, Second Edion
12-6 Chapter 12. Oversaturated Condi ons Operaonal Objecve Type of System Migaon Strategy Definion Maximize Intersec on Throughput Individual Intersec on(s) Split Realloca on Realloca ng split me from under-saturated phases to oversaturated phases. Cycle Length Increase Adding split me to oversaturated phases without reducing split me for minor movements, effec vely increasing the cycle length. Opera on of Closely Spaced Intersec ons on One Controller Opera ng two intersec ons using one controller, allowing close coordina on. Arterial Phase Sequence Modificaon for Le Turns1 Changing the phase sequence to lead or lag le turns, increasing the bandwidth for progressing platoons. Preempon Flushing (âGreen Flushâ) Progressing oversaturated movements by placing calls from downstream intersecons to upstream intersecons. Alternave Timing Plan Flushing Using a ming plan to increase the cycle length and give the majority of the green me to the oversaturated phases. Manage Queues Individual Intersecon(s) Green Extension Adding green me to a phase when a detector exceeds a defined threshold of occupancy. Phase Re-Service Serving an oversaturated phase twice during the same cycle. Phase Truncaon Terminaon of a green interval (despite demand) when there is minimal to no flow over a detector. Arterial Simultaneous Offsets SeÂng offset to zero between intersecons, so that queues move simultaneously. Negave Offsets Starng the green interval earlier at downstream intersecons to allow the downstream queue to dissipate before upstream vehicles arrive. Offsets to Prevent Queue Spillback Choosing offsets that allow the downstream queue to clear before the upstream vehicles reach the end of the downstream queue. Offsets to Prevent Starvaon Choosing offsets that ensure the first released vehicle at the upstream intersecon joins the discharging queue as it begins to move. Network Metering (Gang) Impeding traffic at appropriate upstream points to prevent traffic flow from reaching crical levels at downstream intersecons. Maximize Intersecon Throughput and Manage Queues All Adapve Control Applying detecon data and adapve signal control algorithms to adjust signal ming. Combinaon of Migaon Strategies Combining migaon strategies and/or using them in sequence throughout the oversaturated period. 1 This strategy can also be applied at the individual intersecon level, but is primarily used for arterials. 12.3.1 Migaon Strategies for Individual Intersecons Some strategies are best utilized at individual intersections when mitigating oversaturated conditions. 12.3.1.1 Split Reallocaon When oversaturated conditions exist, it is often necessary to reallocate splits to provide a larger portion of the cycle to oversaturated phases and shorter splits to Exhibit 12-9 Oversaturaon Migaon Strategies Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-7 under-saturated phases. While reallocating split time to accommodate demand on a saturated movement is detrimental to minor movements, the overall beneit is usually recovered through a reduction in delay on the oversaturated approach. Even just a few seconds of reallocated green time can noticeably reduce the growth of overlow queues. By observing the arrival rate and the departure capacity of the oversaturated movement, practitioners can typically calculate the amount of green time required to reduce overlow queues. Phases with lower volume-to-capacity (v/c) ratios are best suited for reallocation of split time, but care should be taken to keep the volume-to- capacity ratios below 1.0. If this is not possible, the phase (or phases) with the most upstream storage space for queued vehicles should be selected for split-time reduction. 12.3.1.2 Cycle Length Increase When split reallocation causes an excessive dis-beneit to minor movements, an increase in cycle length may be required in order to obtain the green time needed for oversaturated movements. This increase does not mean that the cycle length must be long. A common misconception held by many practitioners is that long cycle lengths are needed during oversaturated conditions, which leads to the tendency to increase cycle lengths as queue lengths increase. This misconception is founded on the notion that longer cycle lengths are more eficient because a smaller portion of the cycle is lost due to phase changes. However, research and practical experience have shown that when green intervals are longer than 30 seconds, saturation low rates and overall intersection eficiency decline (1). Therefore, when all approaches at an intersection are oversaturated, the most eficient cycle length is one that terminates service of a phase when the saturation low rate drops, so that subsequent oversaturated phases may be served. Increasing the cycle length at a single intersection tends to disrupt progression at adjacent intersections, so there is a limit to the amount of time that should be added to a cycle. One exception would be when the downstream intersection is capable of handling the increased trafic. In addition to typical cycle length factors (e.g., volume and capacity), oversaturated conditions require consideration of factors such as available storage space, arrival rates on red, and split ratios. A cycle length policy developed by Lieberman, Chang, and Prassas (2) puts an upper bound on cycle lengths to avoid spillback and resulting progression issues. These cycle lengths ensure that the queue formation shockwave dissipates before reaching the upstream intersection. An example based on the Lieberman, Chang, and Prassas policy shows a family of curves (illustrated in Exhibit 12-10) that relate the highest feasible cycle length to speciic link lengths and split ratios. The blue lines in the exhibit are the upper bound for cycle lengths, and the area beneath them represents the cycle lengths that will not generate spillback. For example, a split ratio of 0.5 for the downstream through phase and a 700-foot link length yields a maximum cycle length of 150 seconds before spillback occurs. This graph should be viewed as a rough guide when selecting cycle lengths; many other factors must be considered, including demand, offset, pedestrian requirements, and minimum green times. Substantial increases or decreases in cycle length can cause excessive durations of transition (introduced in Chapter 7), and the effectiveness of interim signal timing parameters during transition periods can be poor. In particular, if the minimum green time for a phase with an overlow queue is reduced during transition, the transition Signal Timing Manual, Second Edion
12-8 Chapter 12. Oversaturated Condi ons period will only worsen conditions. For the mitigation of oversaturated conditions, it is important that any new cycle length is implemented before queues begin to form. 12.3.1.3 Operaon of Closely Spaced Intersecons on One Controller Diamond interchanges are the most common example of locations where more than one intersection may operate jointly on one controller. Such operation permits close coordination between intersections, which can help minimize spillback and starvation, increase system throughput, and manage queues on links with limited storage space. The operation of closely spaced intersections on one controller requires the strategic placement of Âield detection (if actuated), as well as the installation of cabinet wiring speciÂic to this operational structure. Thus, this strategy is not intended to be applied as part of a time-of-day schedule or as an intermittent, reactive solution to trafÂic conditions. If two intersections are fewer than 10 seconds of travel time apart, the use of one controller for intersection operations should be considered. (At a travel speed of 35 miles per hour, this distance is approximately 500 feet.) Most modern controllers can accommodate this type of operation through the use of overlaps and careful setup of phase sequence and splits. Operation of two intersections on one controller may necessitate the use of more than eight phases, four overlaps, and two rings. Several trafÂic control Âirmware programs and state-of-the-art cabinets are available that can accommodate this specialized operation. 12.3.1.4 Green Extension Green extension is a strategy that adds green time to a phase when a detector (or detectors) on an approach exceed some threshold of occupancy. This strategy is often Exhibit 12-10 Example of Cycle Length as a Func on of Split Ra o and Link Distance Signal Timing Manual, Second Edion
12-8 Chapter 12. Oversaturated Condi ons period will only worsen conditions. For the mitigation of oversaturated conditions, it is important that any new cycle length is implemented before queues begin to form. 12.3.1.3 Operaon of Closely Spaced Intersecons on One Controller Diamond interchanges are the most common example of locations where more than one intersection may operate jointly on one controller. Such operation permits close coordination between intersections, which can help minimize spillback and starvation, increase system throughput, and manage queues on links with limited storage space. The operation of closely spaced intersections on one controller requires the strategic placement of Âield detection (if actuated), as well as the installation of cabinet wiring speciÂic to this operational structure. Thus, this strategy is not intended to be applied as part of a time-of-day schedule or as an intermittent, reactive solution to trafÂic conditions. If two intersections are fewer than 10 seconds of travel time apart, the use of one controller for intersection operations should be considered. (At a travel speed of 35 miles per hour, this distance is approximately 500 feet.) Most modern controllers can accommodate this type of operation through the use of overlaps and careful setup of phase sequence and splits. Operation of two intersections on one controller may necessitate the use of more than eight phases, four overlaps, and two rings. Several trafÂic control Âirmware programs and state-of-the-art cabinets are available that can accommodate this specialized operation. 12.3.1.4 Green Extension Green extension is a strategy that adds green time to a phase when a detector (or detectors) on an approach exceed some threshold of occupancy. This strategy is often Exhibit 12-10 Example of Cycle Length as a Func on of Split Ra o and Link Distance Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-9 applied at freeway-to-arterial interchanges where off-ramp queues may affect freeway low and pose safety concerns. It is typically enabled through logic in the local controller that activates an alternative split or, in extreme cases, through a preemption input that immediately serves the oversaturated phase for a predetermined amount of time. Despite its utility, preemption disrupts coordinated operations and is not a recommended strategy to mitigate oversaturation. If the strategy is applied, however, care should be taken to consider the storage space available downstream. If limited storage is available, oversaturation should be addressed across the span of the route as a systemic issue, not as an isolated issue at a single intersection. 12.3.1.5 Phase Re-Service Phase re-service is a strategy that serves a phase twice during the same cycle. Intersections that bene it from phase re-service are often characterized by heavily imbalanced lows on major movements. A commonly implemented phase re-service scenario is one in which a left-turn phase both leads and lags the opposing through movement (illustrated in Exhibit 12-11). In a modern controller, this is con igured through activation of a âphase re-serviceâ or âconditional serviceâ setting. Another scenario in which phase re-service may be effective is when an intersection approach is congested due to the merging of two heavy movements upstream of the intersection. For example, a shared receiving lane for an upstream southbound left-turn movement and eastbound through movement (as shown in Exhibit 12-12) may cause congestion downstream. In this situation, traf ic volumes approach the downstream intersection at two distinct times during the cycle. Excessive delay for one of the two upstream movements may result if the phase at the downstream intersection is only served once per cycle. Exhibit 12-11 Ring-and-Barrier Diagram for Re- Service of Le-Turn Phase Signal Timing Manual, Second Edion
12-10 Chapter 12. Oversaturated Condi ons Another application of the phase re-service strategy is alternating service of the minor movement every other cycle. This strategy may be useful for intersections that have split phasing on the minor street, a constrained cycle length, and are unable to serve both minor street approaches while providing enough time to the major street phases. To alleviate this problem, one minor street approach may be served at the start of the cycle, while the other minor street approach is served toward the cycleâs end. This strategy is also considered to be âdouble cycling,â as all phases are served only once in twice the amount of time as those at adjacent intersections. While there are traditionally eight phases in a dual-ring controller conÂiguration, double cycling requires the use of overlaps. Most modern controllers with sixteen phases and eight overlaps can be conÂigured to provide this type of operation (as illustrated in Exhibit 12-13). 12.3.1.6 Phase Truncaon Phase truncation is a strategy that terminates the green interval (despite demand) when there is minimal to no Âlow over a detector, using logic programmed in the local controller. This strategy may be applied at intersections where there is a lack of downstream capacity or where overÂlow queues for other phases are common, as illustrated in Exhibit 12-14. In the exhibit, northbound left-turning vehicles block the Exhibit 12-12Merging of Two Heavy Upstream Movements Exhibit 12-13 Ring-and-Barrier Diagram for Double Cycling Signal Timing Manual, Second Edion
12-10 Chapter 12. Oversaturated Condi ons Another application of the phase re-service strategy is alternating service of the minor movement every other cycle. This strategy may be useful for intersections that have split phasing on the minor street, a constrained cycle length, and are unable to serve both minor street approaches while providing enough time to the major street phases. To alleviate this problem, one minor street approach may be served at the start of the cycle, while the other minor street approach is served toward the cycleâs end. This strategy is also considered to be âdouble cycling,â as all phases are served only once in twice the amount of time as those at adjacent intersections. While there are traditionally eight phases in a dual-ring controller conÂiguration, double cycling requires the use of overlaps. Most modern controllers with sixteen phases and eight overlaps can be conÂigured to provide this type of operation (as illustrated in Exhibit 12-13). 12.3.1.6 Phase Truncaon Phase truncation is a strategy that terminates the green interval (despite demand) when there is minimal to no Âlow over a detector, using logic programmed in the local controller. This strategy may be applied at intersections where there is a lack of downstream capacity or where overÂlow queues for other phases are common, as illustrated in Exhibit 12-14. In the exhibit, northbound left-turning vehicles block the Exhibit 12-12Merging of Two Heavy Upstream Movements Exhibit 12-13 Ring-and-Barrier Diagram for Double Cycling Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-11 intersection due to downstream congestion. Truncation of the northbound left-turn phase prior to the resultant cross blocking permits other movements to be served while the downstream queue disperses. When a phase is truncated, the spare green time is reallocated to provide extended service to other phases. This type of mitigation is most applicable when queue management is the primary objective and when intermittent oversaturation occurs downstream of the affected phase. Research for NCHRP Project 03-66 (3) concluded that, in an isolated situation, phase truncation can effectively reduce lost green time. However, the effectiveness of truncation depends on the growth and dissipation patterns of the overÂlow queue. For example, if left-turn demand is relatively low and left-turning vehicles do not spillback into the through lane, the overall operational improvements are likely to be minor. 12.3.2 Migaon Strategies for Arterials In most cases, using mitigation strategies for oversaturated arterials in combination with intersection-speciÂic strategies is recommended. 12.3.2.1 Phase Sequence Modifica on for Le Turns Oversaturated conditions can be addressed by modifying the left-turn phase order. In particular, using lead-lag left-turn phasing provides a wider bandwidth for the progression of platoons (introduced in Chapter 7). In oversaturated conditions, lead-lag phasing can help prevent storage bay spillback, storage bay blocking, and starvation. While these principles generally apply to arterial oversaturation, individual intersections and minor movements may beneÂit as well. The alteration of lead-lag settings at an intersection is not a strategy that can typically be applied for short-term, intermittent conditions. When the phase sequence is altered, modern controllers induce transition to reestablish a new offset reference point. If permitted left-turn phasing is used, Âlashing yellow arrow (FYA) operation is necessary to accommodate the phase sequence switch. The overall corridor timing plan, as well as the individual intersection conditions, will dictate which phase to lead and which to lag in a lead-lag sequence (as shown in Exhibit 12-15). If a left-turn movement at a particular approach frequently overÂlows Exhibit 12-14 Phase Trunca on Example Signal Timing Manual, Second Edion
12-12 Chapter 12. Oversaturated Condi ons the available storage bay, the left-turn movement should lead the through movement in order to clear left-turning vehicles. Conversely, a lagging left turn is more appropriate if the through lane typically backs up past the end of the left-turn storage bay. This phase order permits the through vehicles to proceed through the intersection and allows the trapped left-turn vehicles to enter the left-turn lane prior to the beginning of the left- turn phase. For a lagging left turn, additional controller programming is recommended, so that the phase can gap out (and not extend unnecessarily to the end of the coordinated through phase). However, because lagging left turns can use the full split time, it may be beneÂicial to lag the left-turn movement with the higher demand. 12.3.2.2 Flushing Strategies Overly congested, persistent queuing on an oversaturated route is difÂicult to mitigate with measures such as offset adjustment or phase sequence modiÂication. In these extreme conditions, preemption (or âgreen Âlushâ) strategies may be applied after other mitigation strategies have provided little incremental beneÂit. In conditions where minor street volumes compose a notably lower proportion of the total system trafÂic (i.e., 20 percent of total system volume), the effectiveness of Âlushing strategies is substantially greater than that of other mitigation strategy combinations. If minor street volumes compose a larger portion of system trafÂic, however, application of green Âlush strategies will considerably increase minor street delay. Exhibit 12-15 Condi ons for Leading and Lagging Le-Turn Phases Signal Timing Manual, Second Edion
12-12 Chapter 12. Oversaturated Condi ons the available storage bay, the left-turn movement should lead the through movement in order to clear left-turning vehicles. Conversely, a lagging left turn is more appropriate if the through lane typically backs up past the end of the left-turn storage bay. This phase order permits the through vehicles to proceed through the intersection and allows the trapped left-turn vehicles to enter the left-turn lane prior to the beginning of the left- turn phase. For a lagging left turn, additional controller programming is recommended, so that the phase can gap out (and not extend unnecessarily to the end of the coordinated through phase). However, because lagging left turns can use the full split time, it may be beneÂicial to lag the left-turn movement with the higher demand. 12.3.2.2 Flushing Strategies Overly congested, persistent queuing on an oversaturated route is difÂicult to mitigate with measures such as offset adjustment or phase sequence modiÂication. In these extreme conditions, preemption (or âgreen Âlushâ) strategies may be applied after other mitigation strategies have provided little incremental beneÂit. In conditions where minor street volumes compose a notably lower proportion of the total system trafÂic (i.e., 20 percent of total system volume), the effectiveness of Âlushing strategies is substantially greater than that of other mitigation strategy combinations. If minor street volumes compose a larger portion of system trafÂic, however, application of green Âlush strategies will considerably increase minor street delay. Exhibit 12-15 Condi ons for Leading and Lagging Le-Turn Phases Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-13 There are two primary methods used to lush congested corridors. One method utilizes a series of preemption calls to controllers on the route to achieve vehicle progression. The second method initiates a timing plan with a long cycle length that allocates a majority of the split time to the oversaturated route. Either approach can be applied ofline by time-of-day plans, operators at a trafic management center (TMC), or via logic. Operators in Sacramento County, California, for example, manually initiate a route preemption strategy from a TMC based on closed-circuit television (CCTV) surveillance. 12.3.2.2.1 Preempon or âGreen Flushâ In a green lush operation, a series of preemption calls are initiated in sequence from the most downstream location to the most upstream location (as illustrated in Exhibit 12-16). This differs from emergency service preemption, which places calls in the direction of travel. This âbackwardâ sequence aims to clear downstream queues before the release of trafic at upstream signals. For effective implementation of the preemption strategy, the downstream preemption must last as long as necessary for the trafic from upstream locations to progress through the downstream location (or as close as possible, with the 255-second preemption time limitation in most ield controllers). Depending on the persistence of arrival volumes on the oversaturated route, it may be necessary to periodically apply route preemption as queues rebuild during ânormalâ intersection operations. One challenge with using a preemption strategy is the adverse effect of post- preemption transition. Before returning to normal operations, most modern controllers can be conigured to serve the minor street phases immediately after preemption ends. This exit behavior should be carefully considered when applying a green lush strategy. 12.3.2.2.2 Flushing through an Alternave Timing Plan The less severe alternative to the green lush strategy is to implement a timing plan with a long cycle length that allocates a majority of the split time to the oversaturated route. This approach periodically allows a small window of travel opportunity for under-saturated movements, although a majority of the cycle is dedicated to progression on the oversaturated route. This method often results in signiicant queuing on minor streets and under-saturated movements, but can be scheduled as a time-of- Exhibit 12-16 Example Green Flush Sequence Signal Timing Manual, Second Edion
12-14 Chapter 12. Oversaturated Condi ons day pattern if route oversaturation is recurrent and predictable. As shown in Exhibit 12-17, this lush strategy signiicantly improves the travel time of the oversaturated route when compared to less aggressive mitigation strategies. 12.3.2.3 Offset Strategies Offsets establish green bands for eficient progression along an arterial (see Chapter 7 for more information on bandwidth). Most algorithms used to design green bands assume that vehicles can progress along a route unimpeded by overlow queues, and the majority of offset design algorithms assume that queues clear before the irst platoon arrives at the downstream stop bar. When overlow queues begin to form, however, these assumptions are violated, and the practitioner must take a closer look at appropriate offsets. 12.3.2.3.1 Simultaneous Offsets A simultaneous offset is a special case in which the relative offset between two intersections is set to zero (illustrated in Exhibit 12-18). Applications of simultaneous offsets are most effective at locations where the overlow queue on each approach is approximately the same, the green time of each coordinated phase is approximately equal, and, therefore, the queue growth rate is approximately the same. This approach is sometimes referred to as âstore and forward.â Exhibit 12-17 Average Travel Time Using Different Oversatura on Mi ga on Strategies Signal Timing Manual, Second Edion
12-14 Chapter 12. Oversaturated Condi ons day pattern if route oversaturation is recurrent and predictable. As shown in Exhibit 12-17, this lush strategy signiicantly improves the travel time of the oversaturated route when compared to less aggressive mitigation strategies. 12.3.2.3 Offset Strategies Offsets establish green bands for eficient progression along an arterial (see Chapter 7 for more information on bandwidth). Most algorithms used to design green bands assume that vehicles can progress along a route unimpeded by overlow queues, and the majority of offset design algorithms assume that queues clear before the irst platoon arrives at the downstream stop bar. When overlow queues begin to form, however, these assumptions are violated, and the practitioner must take a closer look at appropriate offsets. 12.3.2.3.1 Simultaneous Offsets A simultaneous offset is a special case in which the relative offset between two intersections is set to zero (illustrated in Exhibit 12-18). Applications of simultaneous offsets are most effective at locations where the overlow queue on each approach is approximately the same, the green time of each coordinated phase is approximately equal, and, therefore, the queue growth rate is approximately the same. This approach is sometimes referred to as âstore and forward.â Exhibit 12-17 Average Travel Time Using Different Oversatura on Mi ga on Strategies Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-15 12.3.2.3.2 Negave Offsets Negative offsets are characterized by starting the green interval earlier at the downstream intersection than at the upstream intersection (as illustrated in Exhibit 12-19). This allows the downstream overÂlow queue to disperse before the arrival of the upstream platoon and mitigates wasted green time that results when an upstream platoon arrives at the back of a downstream queue. Generally, the shorter the distance between two intersections, the more critical the design of offsets becomes. Negative offsets are particularly effective with directional trafÂic, while simultaneous offsets may be a more practical solution when congestion exists in both directions. As previously mentioned, more efÂicient operations may be realized with a combination of strategies. Green reallocation and cycle length adjustments coupled with negative (or simultaneous) offsets can improve progression and minimize lost time. TrafÂic demand at downstream intersections can also be reduced by storing trafÂic upstream; metering is discussed in detail in Section 12.3.3.1. Negative offsets alone will not clear an overÂlow queue if the green time is not adequate for the trafÂic demand, but they may reduce the adverse effects of positive progression offsets designed without consideration for downstream queues. Once a queue grows to the extent that all green time for the phase is allocated solely to overÂlow queue clearance (i.e., vehicles in the queue take more than two cycles to clear the intersection), the relationship between offsets is no longer meaningful. Without an increase in green time for the oversaturated phase, adjustments to negative offsets can only marginally reduce the growth rate of overÂlow queues. Once heavy demand rates Exhibit 12-18 Example of Simultaneous Offsets Signal Timing Manual, Second Edion
12-16 Chapter 12. Oversaturated Condi ons subside, however, negative offsets are more ef icient than positive offsets at dispersing over low queues formed during oversaturation. 12.3.2.3.3 Offsets to Prevent Queue Spillback Queue spillback forms when a downstream queue ills the downstream link, preventing upstream vehicles from proceeding through the intersection. When a queue exists at a downstream intersection, vehicles must slow in order to join the back of the queue (as shown in Exhibit 12-20). Particularly for systems with very short link lengths (i.e., in downtown areas), it is important to design the progression strategy appropriately to prevent a ripple effect. Offset settings can help prevent queue growth by providing additional time for a downstream queue to clear before the upstream vehicles reach the end of the downstream queue. Exhibit 12-19 Example of Nega ve Offsets Signal Timing Manual, Second Edion
12-16 Chapter 12. Oversaturated Condi ons subside, however, negative offsets are more ef icient than positive offsets at dispersing over low queues formed during oversaturation. 12.3.2.3.3 Offsets to Prevent Queue Spillback Queue spillback forms when a downstream queue ills the downstream link, preventing upstream vehicles from proceeding through the intersection. When a queue exists at a downstream intersection, vehicles must slow in order to join the back of the queue (as shown in Exhibit 12-20). Particularly for systems with very short link lengths (i.e., in downtown areas), it is important to design the progression strategy appropriately to prevent a ripple effect. Offset settings can help prevent queue growth by providing additional time for a downstream queue to clear before the upstream vehicles reach the end of the downstream queue. Exhibit 12-19 Example of Nega ve Offsets Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-17 12.3.2.3.4 Offsets to Prevent Starva on Starvation occurs when vehicles discharging at an upstream intersection arrive later than the time when the standing queue at a downstream intersection has been discharged. Starvation results in loss of capacity by wasting valuable green time. An Exhibit 12-20 Example of Offset to Prevent Queue Spillback Signal Timing Manual, Second Edion
12-18 Chapter 12. Oversaturated Condi ons ideal offset ensures that the irst released vehicle at the upstream intersection joins the discharging queue at the downstream intersection just as the back of the queue begins to move. Any further delay in the release of vehicles from the upstream location may cause downstream starvation (illustrated in Exhibit 12-21). Exhibit 12-21 Example of Offset to Prevent Starva on Signal Timing Manual, Second Edion
12-18 Chapter 12. Oversaturated Condi ons ideal offset ensures that the irst released vehicle at the upstream intersection joins the discharging queue at the downstream intersection just as the back of the queue begins to move. Any further delay in the release of vehicles from the upstream location may cause downstream starvation (illustrated in Exhibit 12-21). Exhibit 12-21 Example of Offset to Prevent Starva on Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-19 12.3.2.3.5 Combining Offset Strategies Exhibit 12-22 illustrates how various strategies can be used to determine a desirable range of offsets when queues are present. The range of desirable offsets depends on link length, overlow queue length, and queue discharge rates; the resulting feasible zone is deined using relative offsets. For example, a queue ratio of 0.5 (i.e., half of the 800-foot link length is illed with a queue) constrains the feasible offsets between relative negative 30 seconds and 10 seconds. The larger the queue, the larger the need for negative relative offsets to avoid downstream congestion. As discussed previously, a negative relative offset may be infeasible with short block spacing and the need to move congested trafic in both directions. 12.3.3 Migaon Strategies for Networks As oversaturated conditions worsen and affect more intersections and arterials, a more extensive, network-level approach may be necessary. 12.3.3.1 Metering (Gang) Metering (also known as gating) can be applied to impede trafic at appropriate upstream points and to prevent trafic low from reaching critical levels at downstream intersections. Application of the gating strategy is appropriate when downstream spillback is the main cause of congestion. Implementation of this strategy involves identiication of critical intersections in the network that require protection from spillback, as well as identiication of exterior links that can be used to store queues. Metering locations should be determined on a case-by-case basis; general considerations for this selection include ⢠Physical link length and storage capacity, ⢠Locations of trafic generators and trafic attractors, and Exhibit 12-22 Offset Values to Avoid Spillback and Starva on for Intersec ons Spaced at 800 Feet Signal Timing Manual, Second Edion
12-20 Chapter 12. Oversaturated Condi ons ⢠Access to generators and attractors. Exhibit 12-23 illustrates an application of gating. In the exhibit, the depicted oversaturated network can be alleviated with the application of a gating strategy at several points (vehicles shown in red). By reducing green time on those approaches, overÂlow queues can be contained at the network entrance points. However, care must be taken to prevent the oversaturation problem from shifting to another network outside of the consideration area. If appropriately applied, gating can also be used in conjunction with other mitigation strategies to improve performance. Exhibit 12-23 Example of Ga ng Opera on Signal Timing Manual, Second Edion
12-20 Chapter 12. Oversaturated Condi ons ⢠Access to generators and attractors. Exhibit 12-23 illustrates an application of gating. In the exhibit, the depicted oversaturated network can be alleviated with the application of a gating strategy at several points (vehicles shown in red). By reducing green time on those approaches, overÂlow queues can be contained at the network entrance points. However, care must be taken to prevent the oversaturation problem from shifting to another network outside of the consideration area. If appropriately applied, gating can also be used in conjunction with other mitigation strategies to improve performance. Exhibit 12-23 Example of Ga ng Opera on Signal Timing Manual, Second Edion Chapter 12. Oversaturated Condi ons 12-21 During service recovery, gating is typically lifted so that spare network capacity can be used and queues formed outside of the network can be alleviated. If oversaturation is predictable and recurrent, a time-of-day schedule may be developed that applies these strategies during each oversaturation performance period. If the start and end times of network saturation are variable, the decision to use gating may be executed through logic statements or manual control (e.g., TMC operator control). 12.3.4 Adapve Control Adaptive control methods (described in Chapter 9) can improve performance over time-of-day and actuated control operations in under-saturated conditions. While adaptive methods can be effective in delaying the onset of congestion, the ability of adaptive systems to mitigate oversaturated conditions has not been demonstrated. The core features of adaptive control that help to delay the onset of oversaturation include cycle length adjustment, split reallocation, offset adjustment, and phase sequence modiÂication. If congestion is extensive in time and space, the previously described techniques are better able to manage congestion and increase throughput. If congestion is minor, adaptive systems can recover from oversaturation faster than conventional controls. 12.4 REFERENCES 1. Denney, Jr., R. W., L. Head, and K. Spencer. Signal Timing Under Saturated Conditions. Report FHWA-HOP-09-008, Federal Highway Administration, United States Department of Transportation, 2008. 2. Lieberman, E. B., J. Chang, and E. S. Prassas. Formulation of Real-Time Control Policy for Oversaturated Arterials. In Transportation Research Record: Journal of the Transportation Research Board, No. 1727, Transportation Research Board, National Research Council, Washington, D.C., 2000, pp. 77â88. 3. Beaird, S., T. Urbanik, and D. M. Bullock. TrafÂic Signal Phase Truncation in Event of TrafÂic Flow Restriction. In Transportation Research Record: Journal of the Transportation Research Board, No. 1978, Transportation Research Board of the National Academies, Washington, D.C., 2006, pp. 87â94. Adapve control systems should not be expected to eliminate oversaturaon. Rather, these systems are one tool in a box of many other convenonal strategies that address the issue of oversaturated condions. Signal Timing Manual, Second Edion
