Diverging Diamond Interchange Informational Guide, Second Edition (2021)

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.

67 This chapter summarizes considerations and analysis required to develop and verify the lane configuration and phasing scheme for a diverging diamond interchange (DDI) as depicted in the flowchart in Exhibit 5­1. The steps in this chapter will help a practitioner confirm whether a DDI is the most appro­ priate interchange form of the alternatives being considered, and if so, the lane configura­ tion and signal phasing required to provide adequate capacity. This series of steps should be completed before the geometric design is complete, but can be conducted at the same time that an initial horizontal alignment is being developed. 5.1 Preliminary Operations Considerations Because DDIs operate differently than other types of interchanges, this section describes the unique movements at a DDI as well as naming conventions for referencing them. 5.1.1 Directional Naming Conventions Several naming conventions are introduced here to aid in discussion of movements between the DDI and adjacent intersections as illustrated in Exhibit 5­2. The following terms apply to each direction of travel: • Upstream Adjacent Intersection is the last intersection a vehicle would travel through before entering the DDI if traveling along the cross street. • Inbound DDI Crossover is the DDI intersection a vehicle would travel through upon entering the DDI. • Outbound DDI Crossover is the DDI intersection a vehicle would travel through upon leaving the DDI. • Downstream Adjacent Intersection is the first intersection a vehicle would travel through upon leaving the DDI if traveling along the cross street. 5.1.2 Operating Zones DDIs have six types of operating zones that will be referenced throughout this guidebook. The operational effectiveness of each of these zones determines the appropriateness of a DDI. As illustrated in Exhibit 5­3, these operating zones include: 1. Approach Zone—the segment of the cross street between an upstream intersection and the first DDI crossover. This is often referred to as the inbound movement to the DDI. 2. Crossover Zone—the crossover intersection as well as the entry and exit approaches to the zone. C H A P T E R 5 Conceptual Operations

68 Diverging Diamond Interchange Informational Guide SEC 7.6 Exhibit 5-1. DDI operations and design flowchart.

Conceptual Operations 69 3. Bridge Zone—the area between the two crossovers where vehicles will drive over or under the bridge. 4. On-Ramp Zone—the area between the crossover and the freeway merge area. 5. Off-Ramp Zone—the area between the freeway diverge area and the crossover. 6. Departure Zone—the segment of the cross street between the second DDI crossover and a downstream intersection. This is often referred to as the outbound movement to the DDI. By moving traffic to the left side of the street between the crossovers, left­turn movements onto the freeway do not conflict with opposing vehicle traffic (see Exhibit 5­4). As opposed to conventional diamond interchanges where left­turning vehicles are stored between the two ramp terminals, no such turn storage is required at a DDI. 5.1.3 Critical Origin-Destination Movements Exhibit 5­5 introduces a naming convention to clearly define each of the critical origin­ destination movements at a DDI. Through movements on the cross street are shown as north­ bound and southbound movements. This orientation, rather than westbound and eastbound, is arbitrary for the sake of discussion in this chapter, but is how the time­space diagrams will be presented in later sections. Exhibit 5-2. Directional naming conventions. Exhibit 5-3. Operating zones.

70 Diverging Diamond Interchange Informational Guide Exhibit 5-4. Key operational characteristics of a DDI. Exhibit 5-5. Naming convention for DDI movements. 5.2 Signal Timing Conventions Before describing available phasing schemes in detail, this section discusses signal timing basics and conventions used throughout this guidebook. While the examples utilize phasing schemes to illustrate different concepts, details about each of the phasing schemes are not provided until Section 5.3. 5.2.1 Movement Numbering Movements describe user actions at an intersection. Like any intersection, the movements at a DDI (both vehicular and pedestrian) should be assigned numbers that do not change with The exhibits used throughout Section 5.2 are for example purposes only. They are compiled from a variety of phasing schemes (explained in Section 5.3).

Conceptual Operations 71 different geometric designs or signal phasing. Movements are different from both lanes and from phases. Lanes are individually striped areas containing one row of vehicles each (e.g., a through movement could be accommodated by three lanes), while phases are a timing process within the signal controller (e.g., a through and right­turn movement could be assigned to Phase 2). At a DDI, it is common to have 12 one­way vehicular movements (six at each crossover) and eight two­way pedestrian movements (four at each crossover). Choosing a consistent numbering scheme allows for easy referencing of movements, particularly when assigning phases. Exhibit 5­6 shows two examples of movement numbering schemes that can be applied at a DDI (with different pedestrian­facility assumptions). Movement numbers are shown in the gray squares, with a P prefix indicating pedestrian movement. Note that vehicular move­ ments that are traveling in the northbound direction or turn towards the northbound direction are depicted using purple arrows, while the vehicular movements that are traveling in the south­ bound direction or turn towards the southbound direction are depicted using blue arrows. Pedestrian movements are represented with black arrows. For purposes of the operational analysis, this chapter assumes pedestrian facilities are in the median [as shown in Exhibit 5­6(a)]. For DDIs with pedestrian facilities on the outside, pedestrian movements would cross the left­turn and right­turn movements at the ramps [as shown in Exhibit 5­6(b)]. Placing pedestrians in the median or on the outside of the DDI is an important decision that should be made early, as it can affect various aspects of the geometric design and signal timing. Refer to Chapter 3 for more information about the trade­offs of placing pedestrian facilities in the median versus on the outside of the DDI. Movement numbers in these examples are assigned as follows: • Through movements in the approach zones are assigned numbers 2 and 6. • Through movements between the crossovers are assigned numbers 1 and 5. (a) Pedestrian Facility in Median (b) Pedestrian Facility on Outside Exhibit 5-6. Movement numbering scheme.

72 Diverging Diamond Interchange Informational Guide • Right­turn movements from the off­ramps are assigned numbers 4 and 8. • Left­turn movements from the off­ramps are assigned numbers 3 and 7. • On­ramp movements are assigned numbers 9, 10, 11, and 12. • Pedestrian movements are assigned numbers associated with the conflicting vehicular move­ ment. For example, the pedestrian movement that conflicts with Vehicular Movement 10 is assigned number P10. 5.2.2 Phase Numbering Phases (φ) allow one or more movements to be served by a traffic signal controller. Once movements have been assigned to phases, signal timing values (for parameters such as yellow change, red clearance, and minimum green) can be assigned to those phases. At a typical two­ ring, eight­phase intersection, there are conventions for assigning phase numbers (e.g., major street through movements are often assigned Phases 2 and 6; minor street through movements are often assigned Phases 4 and 8). For DDIs within a single jurisdiction, phases should be assigned as consistently as possible. Phase numbering will mostly be dictated by the phasing scheme that is chosen for the DDI. (Phasing­scheme details are summarized in Section 5.3.) There are many overlaps used in the phasing schemes to minimize lost time, particularly related to the clearance required at the crossovers (discussed in detail in Section 5.2.6.1). Overlaps allow movements to operate during more than one phase and should have their own load switches and accompanying signal timing. Overlap lettering can be assigned in a number of ways. However, a practitioner should attempt to assign them in a consistent manner. Phase pairs should be considered for ease of referencing [e.g., northbound through movement assigned to Overlap A (OVA), southbound through movement assigned to Overlap B (OVB)]. In this document, overlap lettering is assigned to movements in the following order: 1. Through and on­ramp movements between the crossovers. 2. Through movements in the approach zones. 3. On­ramp movements in the approach zones. 4. Off­ramp left­turn movements. 5. Off­ramp right­turn movements. Letters are assigned starting with Overlaps A and B (OVA and OVB), moving to Overlaps C and D (OVC and OVD), and so on. Note that overlaps are not used for all movements in every phasing scheme. Exhibit 5­7 is an example of how phases and overlaps can be assigned (using the three­critical­ movement phasing scheme introduced in Section 5.3.2). The phases and overlaps are shown in purple and blue squares for vehicular phases (purple for northbound movements and blue for southbound movements), while the pedestrian phases are shown in black squares. Phase and overlap labels (purple, blue, and black squares) span the movements (gray squares) that are assigned to them. The phases that are included in the overlaps are listed next to each overlap letter. For example, in Exhibit 5­7, Movements 1 and 9 are assigned to OVA, which includes Phases 1 and 4. Movements 2 and 10 are not associated with an overlap, but are instead assigned to Phase 2. It should be noted that some movements may not be signalized based on the design. For example, Movements 9 and 11 are often not signalized. The numbering scheme is designed to accommodate all possible signalization needs.

Conceptual Operations 73 5.2.3 Phases Versus Intervals There are multiple ways that movements at an intersection can be conveyed and represented visually. Two common ways are ring­and­barrier diagrams and interval diagrams. This guide­ book will utilize ring­and­barrier diagrams because it is important that a practitioner understand how movements are programmed in the signal controller. Because interval diagrams have traditionally been utilized for explaining DDI operations, an example of the relationship between a ring­and­barrier diagram and an interval diagram is presented in Exhibit 5­8 (using the three­critical­movement phasing scheme introduced in Section 5.3.2). Each phase shown in the ring­and­barrier diagram represents a certain amount of time. The example shown in Exhibit 5­8 has two rings but no barriers. A new interval begins each time a transition is made between phases. In Exhibit 5­8, there are transitions at the end of Phases 4 and 8, the end of Phase 2, the end of Phase 5, and the end of Phases 1 and 6. The interval diagram is an easy way to represent what is happening at a DDI. However, the diagram is unable to convey the time given to each movement, how that time can shift between phases, and the correlation between the two rings used by the controller. The remainder of this guidebook will use ring­ and­barrier diagrams to explain phasing concepts. 5.2.4 Ring-and-Barrier Diagram Ring­and­barrier diagrams are used throughout the following section to help explain the phasing schemes. For more information on reading a ring­and­barrier diagram, refer to Note: Considerations for overlap delay and extended clearance are summarized in Section 5.2.6.1. Exhibit 5-7. Phase and overlap assignment example (using three-critical-movement phasing scheme).

74 Diverging Diamond Interchange Informational Guide NCHRP Report 812: Signal Timing Manual, 2nd Edition (1). Some considerations for the DDI­ specific ring­and­barrier diagrams are provided in Exhibit 5­9, which uses the four­critical­ movement phasing scheme (Option B) for the example (introduced in Section 5.3.3). Each phase in the ring­and­barrier diagrams is depicted using the green, yellow, and red areas; these areas not only show the phase order but also reflect their approximate duration. Small drawings of a DDI are shown in each phase area to help the reader understand the movements assigned to each phase. Hatching indicates which phases are fixed­time, and labels clarify phase­ length requirements (i.e., based on travel time or clearance of a particular movement). Movements are shown at the appropriate crossover (northern or southern) and are depicted using purple, blue, and black arrows for the northbound vehicular, southbound vehicular, and pedestrian movements, respectively. Overlaps are indicated by a letter next to the associated movement arrow, but labels are also shown above and below the traditional ring­and­barrier diagram to represent the length and order of overlaps. Note: Considerations for overlap delay and extended clearance are summarized in Section 5.2.6.1. Exhibit 5-8. Relationship between ring-and-barrier diagram and interval diagram example (using three-critical-movement phasing scheme).

Conceptual Operations 75 5.2.5 Time-Space Diagram In addition to the phase­assignment and ring­and­barrier diagrams, each phasing scheme described in the next section has a time­space diagram that combines these two elements. For more information about reading a time­space diagram, refer to NCHRP Report 812: Signal Timing Manual, 2nd Edition (1). Some considerations for the DDI­specific time­space diagrams are provided in Exhibit 5­10, which uses the two­critical­movement phasing scheme for the example (introduced in Section 5.3.1). Time is shown on the x­axis, and distance is shown on the y­axis. Ring­and­barrier diagrams are shown for each crossover; they convey the phase order and length and will match the larger ring­and­barrier diagram shown for each phasing scheme. Some phasing schemes use one ring for each crossover (i.e., Ring 1 for the northern crossover and Ring 2 for the southern crossover), and some phasing schemes use both rings at both crossovers. The ring­and­barrier diagrams are separated by the distance between the crossovers. Upstream and downstream signals will use the same distance scale. The bandwidth (i.e., potential for pro­ gression) for each movement is depicted by the purple (northbound) and blue (southbound) bands. The wider the band, the more opportunity for progression between the crossovers for the indicated movement. Use of overlaps results in the bandwidth being shown over the Note: Considerations for overlap delay and extended clearance are summarized in Section 5.2.6.1. Exhibit 5-9. Ring-and-barrier diagram example and conventions (using four- critical-movement phasing scheme Option B).

76 Diverging Diamond Interchange Informational Guide programmed phase clearance times. While the two­critical­movement phasing scheme example only shows bandwidth for the cross street, other phasing schemes in Section 5.3 will depict the available bandwidth for the ramps as well. 5.2.6 Signal Timing Parameter Considerations There are two important distances that influence signal timing at a DDI: (1) distance for a through movement to clear the conflicting off­ramp movement (clearance time) and (2) distance between the crossovers (travel time). Throughout the phasing schemes, these distances will be referenced. 5.2.6.1 Clearance Time Much of the DDI signal timing is impacted by the distance for a through movement to clear the conflicting off­ramp movement. Unlike a conventional intersection, a through movement at a DDI has two distinct clearance times that need to be considered. The first is the time required to clear the opposing through movement at the crossover. This time is typically short because of the width of the crossover. The second is the time required to clear the conflict point with the downstream left­turn or right­turn movement from the off­ramp. This time can be signifi­ cantly longer than the time required to clear the opposing through movement depending on the geometry of the DDI. Exhibit 5-10. Time-space diagram example and conventions (using two- critical-movement phasing scheme).

Conceptual Operations 77 Exhibit 5­11 depicts the total clearance distance for each through movement (approaching and departing) at the crossovers. The options for accommodating the additional clearance time include: • Using short, fixed­time phases following the through movements. • Assigning the off­ramp movements to overlaps, in combination with overlap delay. • Using extended clearance intervals (i.e., longer red clearance) on the through­movement phases (which is least desirable because it is perceived poorly by drivers unless there are site­specific safety concerns due to red­light running). Bicycles generally have slower speeds and acceleration than vehicles, which can impact the time required to clear the conflicting off­ramp movement. If clearance intervals are based on vehicle characteristics, a ramp movement could receive a green indication before a cyclist has cleared the conflict area. Different traffic signal controller vendors have different options for bicycle phasing and timing parameters, but particularly if there are high bicycle volumes at a DDI, the practitioner should consider how to adjust the options above to account for bicycle characteristics. The additional time that is required for a vehicle to clear the off­ramp movements can poten­ tially be utilized by other nonconflicting movements, as is illustrated in Exhibit 5­12. This example makes use of a short, fixed­time phase to accommodate the additional clearance time. The exhibit shows how the northbound through movement (Movement 1 in Phase 4) requires a short clearance time to clear the crossover (shown as time “T1”) and a much longer clearance time to clear the westbound right turn (shown as time “T2”). The exhibit shows the difference between signal timing without (top diagram) and with (bottom diagram) the use of a short, fixed­time phase. The use of this short, fixed­time phase (in combination with overlaps) allows the southbound through movement (Movement 2) to utilize time that would have been lost otherwise. 5.2.6.2 Travel Time The second important distance that impacts DDI signal timing and operations is the distance between the crossovers (depicted in Exhibit 5­13). Depending on the phasing scheme, the time to travel between the crossovers might be used to determine the length of fixed­time phases. Providing a fixed amount of time for a movement to travel the distance between the crossovers can reduce queuing and keep the space between the crossovers clear. Exhibit 5-11. Through-movement clearance considerations.

Note: The rings have been drawn to focus on clearance considerations for the northbound through movement. Not all phases are shown in the ring-and-barrier diagrams. Exhibit 5-12. Impact of crossover clearance time on signal timing. Exhibit 5-13. Travel time between crossovers considerations.

Conceptual Operations 79 5.2.6.3 Cycle Length The travel time from one crossover to the other is also an important factor to consider when selecting the cycle length. Conceptually, if the split for each direction could equal the travel time between crossovers and the cycle length could equal twice that value, progression could be achieved for both directions with an appropriate offset. Exhibit 5­14 illustrates this prin­ ciple using the two­critical­movement phasing scheme, which is explained in more detail in Section 5.3.1. Generally, the travel time between crossovers is too short to result in a viable cycle length, particularly at locations requiring signal timing for pedestrians and bicycles; however, it has been found at several existing DDIs that using a multiple of the travel time for splits and cycle lengths tends to provide reasonably good two­way progression. Because of this finding, a good starting point for determining the cycle length is to measure the travel time between the crossovers and then evaluate the capacity of cycle lengths that are multiples of that number. Since speeds internal to the DDI have been found to be substantially lower than marked speed limits or design speeds, it is important to actually measure the travel time. In addition to internal DDI interactions, cycle lengths should be chosen considering adjacent intersections. More information about adjacent intersections can be found in Section 5.4. Exhibit 5-14. Relationship between splits and travel time (using two-critical- movement phasing scheme).

80 Diverging Diamond Interchange Informational Guide 5.2.7 Phasing Scheme Layout Three phasing schemes will be discussed in this guidebook. Exhibit 5­15 depicts the three pieces of information provided for each phasing scheme, which were described in the previous sections—(1) phase­assignment schematic, (2) ring­and­barrier diagram, and (3) time­space diagram. Each of these pieces relates to the others. The schematic in the top left shows the assign­ ment of movements to phases, which are then depicted in the ring­and­barrier diagram at the bottom. The ring­and­barrier diagram establishes phase order and relative timing for the phases. The ring­and­barrier diagram is then translated for each crossover in the time­space diagram in the top right, which is scaled to match the crossover distance in the phase­assignment schematic. The time­space diagram is ultimately used to identify and compare progression opportunities. 5.3 Phasing Schemes The three phasing schemes discussed in this guidebook were developed to address the following critical­movement scenarios: • Two-critical-movement—emphasizes progression for either the cross­street movements or the off­ramp movements and is most applicable for DDIs with one dominant movement. • Three-critical-movement—emphasizes progression for the cross­street movements and the off­ramp left­turn movements and is most applicable for DDIs with one or multiple domi­ nant movements. The three­critical­movement scheme is often the most flexible and most efficient DDI phasing option. • Four-critical-movement—emphasizes progression for both the cross­street movements and the off­ramp movements and is most applicable to DDIs with low to moderate volumes, either dominant through or left­turn movements and short to medium crossover spacing. The four­critical­movement scheme tends to be less efficient and result in higher delays for DDIs that are approaching capacity, but can be a good option for metering traffic for an adjacent intersection. As the names suggests, the phasing scheme that is selected should be based on the number of critical movements at the DDI. The critical movements will be those that an agency is choosing to prioritize, likely based on demand or other considerations. Additional guidance on selecting Exhibit 5-15. Phasing scheme layout. It is important to note that the phasing schemes are presented assuming a single controller at the DDI.

Conceptual Operations 81 these schemes and a comparison of their operational performance is provided in Section 5.5. Splits in the following phasing scheme examples are for illustrative purposes only; splits should be adjusted to reflect demand. 5.3.1 Two Critical Movements The two­critical­movement phasing scheme can either emphasize progression for the cross­ street movements (depicted in Exhibit 5­16) or the off­ramp movements (depicted in Exhibit 5­17). Considerations for this phasing scheme include the following: • The movements at the northern crossover are assigned to phases in Ring 1, and the move­ ments at the southern crossover are assigned to phases in Ring 2. While set up for a single controller, this convention can also be used if there is a controller at each crossover. • Because each crossover operates in an independent ring, no barriers are required. Coordina­ tion between the crossovers can be achieved through ring offset. Ring offset allows Ring 2 to be offset from Ring 1. The offset for Ring 1 will be referenced to the master clock, while the offset for Ring 2 will be the offset for Ring 1 plus a specified value. Ring offset is not available in every controller, so dummy phases may be required to achieve a similar result. • Exhibit 5­16 depicts a scenario that prioritizes the cross­street movements. The northbound cross­street movements are given time, and then the southbound cross­street movements are given time. Depending on intersection spacing, there is little or no ramp progression with this phasing scheme. • By changing the ring offset, the left­turn movements from the ramps can be prioritized, as shown in Exhibit 5­17. The westbound and eastbound left­turn movements are given time and can immediately proceed through the downstream crossover. Additional time could be given to the left­turn phases if there is demand. Depending on intersection spacing, there is little or no cross­street progression with this phasing scheme. • For the two­critical­movement phasing schemes, the clearance of the through movements is accommodated using short, fixed­time phases. For example, Phase 1 allows Movements 2 and 10 (southbound through and right­turn movements from the cross street) to start before Movement 4 (westbound right­turn movement from the off­ramp). Without this fixed­time phase, the vehicles at the tail end of Movement 1 could still be traveling through the crossover when Movement 4 receives a green. To prevent the fixed­time phases from being skipped, the major phases should be assigned to call them. Note that this additional clearance time could be accommodated using overlap delay or extended clearance intervals (i.e., longer red clearance). See Section 5.2.6.1 for more information on clearance time. • Because pedestrian movements occur in two phases, this phasing scheme results in a high probability that pedestrians will have to wait in the pedestrian island between Movements P1 and P10 at the northern crossover as well as Movements P5 and P12 at the southern cross­ over. Depending on pedestrian volumes, a “Not Ped” feature or other available pedestrian phasing option (such as pedestrian overlaps) could be applied to reduce the likelihood of pedestrians waiting in the island. A “Not Ped” feature would exclude the conflicting vehicle overlap only when there is a pedestrian call. 5.3.2 Three Critical Movements The three­critical­movement phasing scheme can provide progression for the cross­street movements and the off­ramp left­turn movements (depicted in Exhibit 5­18). Considerations for this phasing scheme include the following: • The movements at the northern crossover are assigned to phases in Ring 1, and the move­ ments at the southern crossover are assigned to phases in Ring 2. While set up for a single controller, this convention can also be used if there is a controller at each crossover.

82 Diverging Diamond Interchange Informational Guide Exhibit 5-16. Two-critical-movement phasing scheme (cross-street progression).

Conceptual Operations 83 Exhibit 5-17. Two-critical-movement phasing scheme (ramp progression).

84 Diverging Diamond Interchange Informational Guide Exhibit 5-18. Three-critical-movement phasing scheme.

Conceptual Operations 85 • Because each crossover operates in an independent ring, no barriers are required. Coordina­ tion between the crossovers can be achieved through ring offset. Ring offset allows Ring 2 to be offset from Ring 1. The offset for Ring 1 will be referenced to the master clock, while the offset for Ring 2 will be the offset for Ring 1 plus a specified value. Ring offset is not available in every controller, so dummy phases may be required to achieve a similar result. • Exhibit 5­18 depicts a scenario that gives additional time to Phase 5, so that the southbound cross street and westbound off­ramp movements have larger progression bands. • For the three­critical­movement phasing schemes, the clearance of the through movements is accommodated assuming the use of overlap delay. This places a specified delay on both the left­turn and right­turn movements from the off­ramps. It will allow the through movements from the cross street to clear the ramp entrance areas before the ramp movements receive a green. This additional clearance time can also be achieved through extended clearance (i.e., longer red clearance) on the phases prior to the off­ramp movements. Another alternative is to use short fixed­time phases, as was discussed in the two­critical­movement phasing scheme. See Section 5.2.6.1 for more information on clearance time. • Because pedestrian movements occur in two phases, this phasing scheme results in a high probability that pedestrians will have to wait in the pedestrian island between Movements P1 and P10 at the northern crossover as well as Movements P5 and P12 at the southern cross­ over. Depending on pedestrian volumes, a “Not Ped” feature or other available pedestrian phasing option (such as pedestrian overlaps) could be applied to reduce the likelihood of pedestrians waiting in the island. A “Not Ped” feature would exclude the conflicting vehicle overlap only when there is a pedestrian call. 5.3.3 Four Critical Movements The four­critical­movement phasing scheme can provide progression for the cross­street movements and the off­ramp left­turn movements, while also providing time for the off­ramp right­turn movements (depicted in Exhibit 5­19 and Exhibit 5­20). Considerations for this phasing scheme include the following: • There are two options for this phasing scheme—Option A and Option B. Option A (shown in Exhibit 5­19) prioritizes the off­ramp left­turn and right­turn movements. Option B (shown in Exhibit 5­20) prioritizes the cross­street movements. Both options have phases based on travel times between the crossovers. The geometry of the crossovers and distance between them will play a critical role in the signal timing for the four­critical­movement phasing scheme. • Unlike the two­critical­movement and three­critical­movement phasing schemes, the four­ critical­movement phasing scheme does not separate the crossovers into separate rings. With Option B, the phases at the northern and southern crossovers could easily be separated into two rings, but by not separating them, the designer can easily compare Option A and Option B. • Phases 1 and 5 are used as “dummy” phases, with no movements assigned to them. This provides consistency for numbering between the alternatives. • For the four­critical­movement phasing schemes, the clearance of the through movements is accommodated assuming the use of overlap delay. This places a specified delay on both the left­turn and right­turn movements from the off­ramps. It will allow the through movements from the cross street to clear the ramp entrance areas before the ramp movements receive a green. This additional clearance time can also be achieved through extended clearance (i.e., longer red clearance) on the phases prior to the off­ramp movements. Another alternative is to use short, fixed­time phases, as was discussed in the two­critical­movement phasing scheme. See Section 5.2.6.1 for more information on clearance time.

86 Diverging Diamond Interchange Informational Guide Exhibit 5-19. Four-critical-movement phasing scheme Option A.

Conceptual Operations 87 Exhibit 5-20. Four-critical-movement phasing scheme Option B.

88 Diverging Diamond Interchange Informational Guide • Both options use fixed­time travel­time phases for efficiency (unrelated to clearance for the ramps). – In Option A, the fixed­time phases are based on the travel time for Movements 2 and 6, which are cross­street through movements. To prevent vehicles at the tail end of those movements from getting stuck between the crossovers, these phases should be based on the time a vehicle needs to travel between the crossovers. – In Option B, the fixed­time phases are based on the travel time for Movements 3 and 7, which are off­ramp left­turn movements. This option will operate most efficiently if the fixed­time phases provide enough time for the off­ramp left­turn movements to travel to the downstream crossover, so that the downstream crossover turns green as the first vehicles from Movements 3 and 7 reach the intersection. • Option B has another set of phases that, while not fixed­time, are dependent on travel time. Phases 4 and 8 should be at least long enough for Movements 2 and 6 to reach the down­ stream crossover. Much like the fixed­time phases in Option A, vehicles at the tail end of those movements should not get stuck between the crossovers. Depending on the time required to finish serving Movements 3 and 7 (off­ramp left­turn movements), the phases may need to be longer than the required travel time. • Because pedestrian movements occur in two phases, this phasing scheme results in a high probability that pedestrians will have to wait in the pedestrian island between Movements P1 and P10 at the northern crossover as well as Movements P5 and P12 at the southern cross­ over. Depending on pedestrian volumes, a “Not Ped” feature or other available pedestrian phasing option (such as pedestrian overlaps) could be applied to reduce the likelihood of pedestrians waiting in the island. A “Not Ped” feature would exclude the conflicting vehicle overlap only when there is a pedestrian call. 5.3.4 Specialized Signal Timing Applications The discussion in the previous section introduced three basic phasing schemes. Many variations of these schemes exist that may alter the phase sequence or duration of phases. It is expected that customization may be needed to serve site­specific needs. Examples of specialized signal timing that can enhance DDI operations for specific applications include: • Pre-timed control—can be used as an alternative to the typically­recommended actuated control at DDIs. • Half cycle—can be used to accommodate multiple platoons from an upstream adjacent signalized intersection. • Vehicle preemption (“ramp flush”)—allows both left­turn and right­turn movements from the freeway to be served concurrently to help reduce spillback onto the freeway mainline. • Dynamic overlap phasing—can be used to mitigate clearance time conflicts. • Meter traffic at upstream adjacent signalized intersection—can be used to minimize queue spillback into the DDI. • Exclusive pedestrian phase—can be used to reduce delay for pedestrians crossing vehicular movements to and from the freeway. • Transit preferential treatment—particularly for DDIs with center­running light rail or a center­running transitway. 5.3.4.1 Pre-Timed Control Actuated signal control generally provides greater flexibility and more efficient service for roadway users, and in most cases, coordination can still be achieved between the crossovers when using actuated control. However, in some applications with two controllers, pre­timed control may be an option to maintain the time­space relationship between the intersections.

Conceptual Operations 89 With actuated­coordinated control, minor movements can gap out or be skipped if demand is low. The extra time is often given to the coordinated movements. Depending on the phasing scheme that is chosen, there may be less added benefit for actuated control because having a movement gap out at one crossover may not greatly benefit progression at the other. A practi­ tioner should also consider time­of­day effects. For example, pre­timed control may be advanta­ geous during the peaks, and actuated control may be advantageous during off­peak times. 5.3.4.2 Half Cycle Since DDI signals generally have fewer signal phases than other intersections on a given corridor, it may frequently be beneficial to run the crossovers at half of the cycle length of the cross street. (It is also possible to run multiple cycle lengths at the DDI, provided the sum of those cycle lengths equals the cycle length of the overall corridor.) Using a half cycle is particularly appealing when there are two heavy upstream movements, such as a through move­ ment and a heavy side­street left­turn movement. If the DDI signals are running a half cycle, both movements have a chance of being progressed through the interchange. Half cycling can also be applied when the adjacent intersection uses split phasing for the side­street move­ ments. When selecting a cycle length, it is critical to consider progression along the entire corridor. For more information about corridor considerations, refer to Section 5.4.1. Another option is to run just one crossover at a half cycle when there is a heavy through move­ ment and a heavy left­turn movement from one of the ramps. In that circumstance, it would be possible to progress both the upstream through movement and the off­ramp left­turn move­ ment through the adjacent crossover. The trade­off in this case is that in every other cycle, the opposing through movement would be stopped internal to the DDI; this application works best with very directional movements. 5.3.4.3 Vehicle Preemption (“Ramp Flush”) Preemption can provide a good mechanism for alleviating excessive queuing (i.e., by “flushing” vehicles through an intersection); however, it should be used sparingly and only when a safety concern cannot be addressed by some other means. Two safety­related reasons to use preemption at a DDI include emergency vehicles and queue spillback onto the freeway. Preemption could be used by emergency responders where traffic demand prevents accept­ able emergency response times. Preemption in this case would flush vehicles through an inter­ section in the emergency vehicle’s direction of travel, allowing emergency vehicles to respond more quickly to emergencies. At freeway exit ramps with high demand from the left­turn or right­turn movements, pre­ emption can be used to help flush queues that are on the verge of spilling back onto the freeway. A detector is placed on the freeway off­ramp, between the exit ramp gore and the DDI signal to detect the presence of excessive queues that have the potential to spill back onto the freeway. When the queue extends over the loop for a specified amount of time, the signal is preempted to call a dedicated phase for the left­ and right­turn movements from the ramp, as illustrated in Exhibit 5­21. The placement of the queue detector is a function of observed queuing patterns and should be placed far enough upstream of the DDI as to not get triggered every cycle (i.e., greater than the average queue) but still downstream of the ramp gore on the off­ramp itself. This ramp preemption method is used regularly by several states. While preemption is capable of alleviating excessive queuing, it should be used sparingly. It should be applied at locations where other signal timing or design strategies (e.g., longer ramps or more lanes) are infeasible. Preemption of a signal on a recurring basis could cause excessive delays and queues to form at intersections as additional phases are called or split times are increased, knocking the controller out of coordination for several cycles.

90 Diverging Diamond Interchange Informational Guide 5.3.4.4 Dynamic Overlap Phasing As discussed in Section 5.2.6, additional clearance time is needed at DDIs to ensure that cross­street traffic can clear conflict points with left turns and right turns from the freeway. Those clearance times are estimated based on vehicular travel times. In some cases, those assumed travel times may be too conservative, but in others they may not provide sufficient time (e.g., in the case of slow­moving vehicles). Agencies can mitigate potential conflicts through dynamic overlap phasing. Dynamic overlap phasing uses detection to extend the overlap phases (being used for clear­ ance) in the case of slow­moving vehicles (such as bicycles) on the cross street. In other words, if a slow­moving vehicle is detected as it approaches the conflict point, the overlap phase can extend to ensure that the phase does not end prior to the vehicle clearing the freeway merge point. Alternatively, agencies can use supplemental signals (and stop bars) to stop cross­street traffic between the crossover and the freeway merge point. This strategy should only be applied when there is sufficient storage for one to two vehicles (per lane) without having them spill back into the crossover. It is only appropriate for DDIs with sufficient spacing between the crossover and the freeway merge point. More information about this treatment can be found in Chapter 7. 5.3.4.5 Meter Traffic at Upstream Adjacent Signalized Intersection A downstream adjacent signalized intersection should be optimized so that throughput can be maximized with minimal queue spillback into the DDI. If spillback still occurs, the throughput of the upstream adjacent signalized intersection could be artificially reduced to match the Exhibit 5-21. Ramp flush.

Conceptual Operations 91 capacity of the downstream signal. The queue that was spilling back from the downstream adjacent intersection into the outbound crossover would then be transferred to the upstream adjacent intersection. This would reduce the chance for queue spillback through the outbound DDI movement. The upstream adjacent signalized intersection should be looked at carefully to make sure that demand starvation is not occurring at the inbound DDI crossover. The DDI crossover is expected to provide more throughput than the upstream adjacent signalized intersection. This creates an opportunity to use additional green time at the DDI to progress left­ and right­turn movements from the side street at the upstream intersection. 5.3.4.6 Exclusive Pedestrian Phase The phasing schemes presented in this chapter assume that all pedestrian movements are signalized and run concurrently with nonconflicting vehicular movements. In the phasing schemes, two pedestrian movements (one at each crossover) occur in multiple stages (i.e., pedes­ trians cross to an island and wait). Depending on pedestrian volumes, a “Not Ped” feature or other available pedestrian phasing option (such as pedestrian overlaps) could be applied to reduce the likelihood of pedestrians waiting in the island. However, another option to reduce pedestrian delay is an exclusive pedestrian phase. Use of an exclusive pedestrian phase would create additional delay for other roadway users, but would allow pedestrians to complete all movements during a single phase. 5.3.4.7 Transit Preferential Treatment Some DDIs have center­running light­rail or a center­running transitway. Transit preferential treatment can be used to accommodate such arrangements. Often, an exclusive transit phase is applied that stops the cross­street vehicular movements but can allow left and right turns from the freeway ramps. Depending on traffic patterns, holding the left turns on the ramps may prevent significant queuing between the crossovers. Preferential treatment will require site­specific adjustments. 5.3.5 Understanding the Benefits and Challenges of DDI Phasing Schemes The three basic DDI phasing schemes each have their unique benefits as well as challenges. Recommendations for which phasing scheme is most appropriate for a specific DDI depend on many factors, including DDI geometry, traffic volumes and patterns, need for specialized signal timing applications, and ultimately, agency preferences. In the sections below, dominant movements refer to those identified in Section 5.4.3. 5.3.5.1 Two-Critical-Movement Phasing Scheme Benefits and Challenges Exhibit 5­22 summarizes the benefits and challenges of the two­critical­movement phasing scheme. In summary, the two­critical­movement phasing scheme is high efficiency and high capacity; it is simple to understand and implement, and it is adaptable to various crossover spacings. However, it has limited ability to progress multiple dominant movements. It is recommended for DDIs with high volume­to­capacity ratios and one dominant movement. 5.3.5.2 Three-Critical-Movement Phasing Scheme Benefits and Challenges Exhibit 5­23 summarizes the benefits and challenges of the three­critical­movement phasing scheme. The three­critical­movement phasing scheme is high efficiency and high capacity; it can progress multiple dominant movements, and it is adaptable to various crossover spacings. However, it may result in internal stops for nondominant movements. It is recommended as The three-critical-movement phasing scheme is recommended as the default when starting an analysis.

92 Diverging Diamond Interchange Informational Guide the default DDI phasing scheme, especially for DDIs with high volume­to­capacity ratios and multiple dominant movements. 5.3.5.3 Four-Critical-Movement Phasing Scheme Benefits and Challenges Exhibit 5­24 summarizes the benefits and challenges of the four­critical­movement phasing scheme. The four­critical­movement phasing scheme is the least efficient phasing scheme, but is capable of progressing all movements and eliminating internal queues at the DDI. The scheme is limited in applicability to DDIs with one or more high volume­to­capacity ratio movements. It is recommended for DDIs with low to moderate volumes and short to medium crossover spacings. 5.4 System Needs One of the most critical operational considerations in the design and implementation of a DDI is how to integrate the interchange with adjacent intersections, which requires site­ specific investigation. Most DDIs are likely to be constructed in developed areas (or soon­to­be developed areas). In these cases, the cross­street portion of the DDI serves as an access route to various land uses in the vicinity of the interchange. The adjacent intersections and access Benefits Challenges + Ability to coordinate through movement on the cross-street or dominant left-turn movement from the ramps – Limited ability to progress multiple movements (e.g., both cross street and movements from the ramps) + Generally easy to understand/implement and troubleshoot in the field due to low complexity of phase assignments – May result in more stops internal to the DDI than other strategies + Minimizes lost time because of the low number of phases + Highest potential capacity of the three phasing schemes + Adaptable to any crossover spacing Exhibit 5-22. Benefits and challenges of the two-critical-movement phasing scheme. Benefits Challenges + Ability to coordinate through movements on the cross-street and left-turn movements from the ramps – More complex than two-critical-movement phasing scheme + Possible to troubleshoot in the field due to the low complexity of phase assignments – Less efficient than two-critical-movement phasing scheme + Moderate lost time with only three critical phases – May result in stops internal to DDI for nondominant movements + High-capacity phasing scheme for multiple dominant movements + Adaptable to any crossover spacing Exhibit 5-23. Benefits and challenges of the three-critical-movement phasing scheme.

Conceptual Operations 93 points may be closely spaced, resulting in significant interaction between the interchange and adjacent intersections. Adjacent intersections may also be roundabouts or have alternative intersection forms. As with any interchange project, the practitioner should carefully consider existing land uses as well as planned and proposed developments that are expected to impact future growth of traffic through the interchange. Having a good understanding of the traffic volume levels and traffic patterns through the interchange is critical to properly design the DDI geometry and signal timing. For retrofitting existing interchanges, the practitioner should pay close atten­ tion to potential volume metering taking place for congested operations. In other words, the practitioner needs to assure that a valid estimate of the true demand at the interchange is used to design and time the DDI because the volumes served by an existing (congested) interchange may be less than those demands. Field observations at various DDIs in the United States have shown that the observed delays and queues at a DDI are often due to queue spillback effects and capacity constraints at down­ stream signalized intersections (2). It is therefore critical that these adjacent intersections are explicitly considered prior to the design of a DDI. There are two primary challenges of integrating a DDI into a corridor or system. The first is that the cross­street through movements cannot be processed simultaneously at the cross­ overs. Instead, each crossover has to process the cross­street through traffic in an alternating sequence, which impacts the ability to progress movements through the interchange in both directions. Two­way cross­street coordination at a DDI is possible, but progression opportuni­ ties will depend heavily on the phasing scheme at the DDI as well as the spacing of the adjacent intersections. The second challenge is that there are typically fewer conflicting movements at the DDI crossovers than at the adjacent intersections (e.g., left­turn movements, minor street movements requiring split phasing). This makes progression between the adjacent intersections and the DDI challenging. 5.4.1 DDI Corridor Signal Timing Three basic DDI phasing schemes were introduced in Section 5.3, including two critical move­ ment, three critical movement, and four critical movement (Option A and Option B). This section discusses these phasing schemes in a corridor context. Exhibit 5­25 through Exhibit 5­28 expand the time­space diagrams for the three phasing schemes to include upstream and Benefits Challenges + Ability to progress all movements through the DDI – Works best with balanced volumes and may be challenging with one or more dominant movements + Minimizes stops internal to the DDI (resulting in a better user experience) – More difficult to understand/implement and troubleshoot in the field due to complexity of phase assignments + Most flexible and adaptable phasing scheme – Highest lost time among the three phasing schemes because of the number of phases – Less capacity than other phasing schemes – Inefficient for wide crossover spacings Exhibit 5-24. Benefits and challenges of the four-critical-movement phasing scheme. Exhibits 5-25 through 5-28 are not intended to provide generalized guidance for DDI corridor performance, but rather to illustrate how to consider and evaluate DDIs in a corridor context. The specific coordination patterns and bandwidths shown will change with different sets of assumptions.

94 Diverging Diamond Interchange Informational Guide Exhibit 5-25. Two-critical-movement phasing in corridor context.

Conceptual Operations 95 Exhibit 5-26. Three-critical-movement phasing in corridor context.

96 Diverging Diamond Interchange Informational Guide Exhibit 5-27. Four-critical-movement phasing (Option A) in corridor context.

Conceptual Operations 97 Exhibit 5-28. Four-critical-movement phasing (Option B) in corridor context.

98 Diverging Diamond Interchange Informational Guide downstream adjacent intersections. The use of time­space diagrams is strongly encouraged to visualize corridor bandwidth and properly evaluate the selected phasing scheme in a corridor context. It is important to note that these diagrams illustrate potential progression opportunities but do not illustrate queuing impacts. When interpreting these exhibits, a practitioner should understand the various assumptions that had to be made to create them, including: 1. Splits and phase sequence at the DDI and adjacent intersections, 2. Spacing of the DDI crossovers and adjacent intersections along the cross street, and 3. Speeds along the cross street (inside and outside the DDI). For purposes of demonstration, the upstream and downstream adjacent intersections in Exhibit 5­25 through Exhibit 5­28 use eight phases in two rings. While the offsets for the upstream and downstream intersections have been optimized for cross­street progression, the splits and phase sequence at the adjacent intersections are the same for all of the scenarios. This allows a direct comparison of the three phasing schemes. It should be noted that, in these examples, the upstream adjacent intersection, downstream adjacent intersection, and DDI crossovers are equally spaced. This is the ideal scenario for progressing cross­street traffic and was chosen for illustrative purposes only. Other spacing scenarios will likely result in smaller bands along the corridor. The purple (northbound) and blue (southbound) areas represent the bandwidth available based on the entire corridor (i.e., upstream and downstream adjacent intersections included). The reader may notice that the total available bandwidth is relatively similar for all three phasing schemes when considering the corridor context. However, the phasing scheme that is chosen at the DDI influences which movements receive the progression opportunities. With a limited amount of available bandwidth, an agency will have to choose priorities at the DDI. 5.4.2 Corridor Operational Considerations and Common Challenges Building on the general coordination discussion, the following items should be considered in the process of integrating a DDI into a signal system. This section presents operational consid­ erations for timing DDIs in a corridor context and describes some common operational chal­ lenges observed along DDI corridors. Section 5.4.4 then presents several strategies that can be employed to address these challenges. 5.4.2.1 Crossover Spacing DDI crossover spacing is a key factor when selecting a phasing scheme. In general, shorter crossover spacing provides less queue storage internal to the DDI but also less time required to clear the interchange. Longer crossover spacing provides the ability to match the travel time between the crossovers and the upstream and downstream adjacent intersections. However, longer distances can be inefficient for clearing traffic between the crossovers. 5.4.2.2 Speed Profiles Free­flow speeds along a corridor are limited by the geometrics of the DDI. Field studies at DDIs in the United States have shown that free­flow speeds through and between the crossovers are lower than the posted speed limit, even without interaction effects of other traffic (2). Free­ flow speeds for the left­turn and right­turn movements are also limited by geometry. Field free­ flow speeds from seven DDIs are summarized in Exhibit 5­29. The figure shows the field­observed mean speeds relative to the posted speed limit. For each observed mean speed, a range of speeds plus/minus one standard deviation is shown as well. Exhibit 5­29 suggests that the geometric design of the DDI controls the crossover free­flow speeds, independent of the speed limit.

Conceptual Operations 99 0 5 10 15 20 25 30 35 40 45 50 Sp ee d (M ile s Pe r H ou r) Speed Limit Speed between Crossovers Crossover Speed Left-Turn Speed from Freeway Exhibit 5-29. Field-measured speeds (mean plus/minus one standard deviation) at DDI sites. Free­flow speeds impact the coordination speed of the DDI with adjacent intersections and potentially the capacity of traffic movements. At the same time, a traffic­calming benefit is pro­ vided that can be important for accommodating nonmotorized users. While it is unclear if DDI speeds will always be below the speed limit or if drivers will adjust over time, the speed profiles are an important consideration when evaluating a DDI in a corridor context. They are also useful for practitioners conducting simulation analyses of newly­designed facilities. 5.4.2.3 Queue Spillback Queue spillback may occur in the departure zone downstream of a DDI. If a DDI crossover signal has a potential discharge volume that exceeds the processing capacity of the downstream adjacent signalized intersection, queues can spillback into the DDI, affecting outbound move­ ments as shown in Exhibit 5­30 and Exhibit 5­31. Source: Field Evaluation of Double Crossover Diamond Interchanges (2). Exhibit 5-30. Queue spillback into DDI from downstream adjacent signal at Harrodsburg Road in Lexington, Kentucky.

100 Diverging Diamond Interchange Informational Guide Exhibit 5-31. Queue spillback at a DDI. Exhibit 5-32. Demand starvation at a DDI. 5.4.2.4 Demand Starvation Demand starvation occurs when a signal phase is green and no traffic is present. This can occur when traffic is held at an upstream adjacent signalized intersection along a corridor. Demand starvation is most likely to occur at the inbound approach to a DDI, where the capacity of the first DDI crossover may be greater than the capacity of the upstream adjacent intersection. When demand starvation occurs, the throughput of the DDI movement will be less than its capacity. Demand starvation is illustrated in Exhibit 5­32. 5.4.2.5 Lane Utilization Interchanges often have imbalanced lane utilization because drivers making turning move­ ments tend to pre­position in advance of the interchange. Field observations show that pre­ positioning is more likely to occur at a DDI compared to a conventional diamond interchange (2). Some of the lane imbalance is likely attributable to the lane configuration at the second cross­ over. In Exhibit 5­33, the three­lane road has a shared left­through lane at the second crossover, whereas in Exhibit 5­34, the three­lane road has an exclusive left­turn lane. The shared­lane configuration is expected to result in a mix of left­turn and through traffic in the leftmost lane at the first crossover, while the exclusive lane is expected to limit the left lane to left­turn traffic only. Lane utilization at the first crossover may be imbalanced in both cases as a function of the amount of left­turning traffic in the traffic stream. Field observations indicate that left­

Conceptual Operations 101 lane utilization can be predicted by the left­turn­demand ratio (2, 3, 4). Lane­utilization models are integrated in the Highway Capacity Manual (HCM6) (5) analysis method for DDIs. The models predict lane utilization for the inbound movement at DDI crossovers as a function of lane geometry and the left­turn demand ratio. The reader is referred to the HCM6 for additional details on these models and how to use them. 5.4.2.6 Weaving Maneuvers Weaving maneuvers result when an adjacent intersection is in close proximity to a DDI, as illustrated in Exhibit 5­35. Many DDIs prohibit right­turn­on­red (RTOR) maneuvers from the freeway exit ramp. This has the benefit of reducing the number of conflicts associated with weaving maneuvers. For DDIs with RTOR allowed or for DDIs with a continuous lane addition serving the right turn, weaving can significantly impact corridor operations and safety. Besides RTOR restrictions, weaving conflicts can be minimized by lengthening the distance to the adjacent intersection. 5.4.3 Assessing Signal Phasing Schemes in a Corridor Context For any operational analysis methodology, a practitioner must first select a phasing scheme to evaluate. This section provides guidance on selecting an initial phasing scheme to test based on the concept of dominant movements. Exhibit 5­36 illustrates eight common dominant­ movement schemes that exist at DDIs. Before proceeding with choosing a phasing scheme, Exhibit 5-33. Shared left-through lane at a DDI. Exhibit 5-34. Left-turn-only lane at a DDI.

102 Diverging Diamond Interchange Informational Guide a practitioner must determine which dominant­movement patterns are likely to occur at the DDI in question, likely through a planning­level demand assessment. The dominant­movement schemes may vary based on time of day. The various phasing schemes described in Section 5.3 can all provide acceptable perfor­ mance, but how a DDI will operate under a particular phasing scheme significantly depends on the crossover spacing distance and traffic patterns. The following guidance provides informa­ tion about the level of delay and queuing that might be expected when using a specific phasing scheme (i.e., two, three, or four critical movement) for a particular crossover spacing distance and set of dominant movements (i.e., one of the eight dominant­movement patterns). Exhibit 5­37 illustrates how the movements are numbered for queuing purposes. The move­ ments listed as potentially having significant queues are those that are anticipated to experience Exhibit 5-35. Weaving maneuver and conflict points for DDI right turn from freeway. Exhibit 5-36. Dominant-movement patterns.

Conceptual Operations 103 long queues. For example, queues for Movements 1 and 5 are considered to be significant if they extend beyond the crossover spacing (300 feet, 750 feet, and 1,200 feet, respectively). A practitioner should assess the list of significant queue movements considering the site in question. If geometry makes it impossible to change the lane configuration at a ramp approach, for example, a long off­ramp queue might indicate that a different phasing scheme should be tested. It is important to emphasize that the following tables assume that the dominant movements are at or near capacity. Clearly, a geometric modification such as a larger cross­section with additional lanes would also be likely to achieve acceptable operations. Under low to moderate volumes, any of the three phasing schemes could result in acceptable performance, although the phasing schemes could produce different results from each other. The exhibits and tables in this section are simply meant to help a practitioner choose an initial phasing scheme to test as he or she initializes an operational analysis (using the tools explained throughout the remainder of this chapter). Exhibit 5­38, Exhibit 5­39, and Exhibit 5­40 (for crossover spacing distances of 300 feet, 750 feet, and 1,200 feet, respectively) provide a starting point for estimating the anticipated Exhibit 5-37. Significant-queue- length movement numbering. Note that the following tables assume the dominant movements are at or near capacity.

104 Diverging Diamond Interchange Informational Guide Heavy Demand Path(s) (Reference Exhibit 5-36) Two-Critical-Movement Phasing Three-Critical- Movement Phasing Four-Critical- Movement Phasing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Through Low - Low - Medium 6, A6 Left Off Medium 1, 5, 6, 7, 8, A6 Medium 4, 7, 8 Medium 3, 4, 6, 7, 8, A6 Right Off Low 4 Low 3, 4, 7, 8 High 3, 4, 6, 7, 8, A6 Left On Medium 3, 4, 5 Medium 3, 4, 5 Medium 3, 4, 5, 7 Through + Left Off Low 1, 7, 8 Low 7, 8 Medium 3, 4, 6, 7, 8, A6 Through + Right Off Low 4 Low 3, 4, 7, 8 Medium 3, 4, 6, 7, 8, A6 Through + Left Off + Right Off Low 1, 4, 7 Low 3, 4, 7, 8 Medium 3, 4, 6, 7, 8, A6 Balanced Low - Low - Low - Exhibit 5-38. Anticipated phasing-scheme queuing for a crossover spacing of 300 feet. Heavy Demand Path(s) (Reference Exhibit 5-36) Two-Critical-Movement Phasing Three-Critical- Movement Phasing Four-Critical-Movement Phasing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Through Low - Low - Medium 3, 4, 6, 7 Left Off Low 7, 8 Low 7 Medium 3, 4, 6, 7, 8 Right Off Low 4 Medium 3, 4, 7, 8 Medium 3, 4, 5, 7, 8 Left On Low - Medium - Medium 3, 4 Through + Left Off Low 7 Low 7 Medium 3, 4, 6, 7, 8, A6 Through + Right Off Low 4 Low 3, 4, 7, 8 Medium 3, 4, 7, 8 Through + Left Off + Right Off Low 4, 7 Low 3, 4, 7, 8 Medium 3, 4, 6, 7, 8, A6 Balanced Low - Low 4, 7 Low 3, 4 Exhibit 5-39. Anticipated phasing-scheme queuing for a crossover spacing of 750 feet.

Conceptual Operations 105 delay impacts of the three phasing schemes (two, three, and four critical movements). An analyst should first use these exhibits to determine which phasing scheme(s) may be most appropriate for a given scenario. Once one or more schemes are chosen, the analyst should consider the queue interactions for critical approaches, which are also summarized in the exhibits. The interpretation of Exhibit 5­38, Exhibit 5­39, and Exhibit 5­40 is as follows: • Delay rating of “low” means that the dominant movements are readily accommodated under the given scenario. However, an analyst should still pay attention to any movements that are anticipated to experience significant queuing. • Delay rating of “medium” means that the dominant movements may be accommodated under the given scenario, but operations should be evaluated carefully with the chosen operational analysis tool. • Delay rating of “high” means that it will likely be difficult to accommodate the dominant movements under the given scenario. If other phasing schemes are anticipated to have lower delay for the given scenario, the analyst should consider testing them first. 5.4.4 Strategies to Improve Corridor Operations This section discusses strategies to help with corridor operations involving a DDI. It should be noted that most of the strategies are focused on maximizing vehicle throughput at the DDI and adjacent signalized intersections. Several strategies can be applied at the DDI crossovers: • Half cycle, • Signalized on­ramp left turn, • Dedicated phase for concurrent off­ramp left and right turns, • LTOR or RTOR allowed at off­ramp, and • Dynamic overlap phasing. Heavy Demand Path(s) (Reference Exhibit 5-36) Two-Critical-Movement Phasing Three-Critical- Movement Phasing Four-Critical-Movement Phasing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Anticip. Level of Delay Movements that May Experience Significant Queuing Through Low - Low - Medium 3, 4, 6, A6 Left Off Low 7 Low 7 High 3, 4, 6, 7, 8, A6 Right Off Low 4 Low 4, 3 High 3, 4, 6, 8, A6 Left On Low - Medium - High 3, 4, 6, 8, A6 Through + Left Off Low 7 Low 7 Medium 3, 4, 6, 7, 8, A6 Through + Right Off Low 4 Low 4 Medium 3, 4, 6, 7, 8, A6 Through + Left Off + Right Off Low 4, 7 Low 4 Medium 3, 4, 6, 7, 8, A6 Balanced Low - Low - Medium 3, 4, 8 Exhibit 5-40. Anticipated phasing-scheme queuing for a crossover spacing of 1,200 feet. Trade-offs between vehicles and other modes should be considered for all of the listed strategies.

106 Diverging Diamond Interchange Informational Guide Other strategies can be applied at the adjacent signalized intersections in order to influence operations at the DDI: • Optimize timing and/or meter traffic at upstream adjacent signalized intersection, • Alternate side­street phases at downstream adjacent signalized intersection, • Lead/lag phasing for outbound left turns at downstream adjacent signalized intersection, and • Eliminate phases at adjacent signalized intersection. Additionally, one strategy can be applied at the corridor level (i.e., at the DDI crossovers and adjacent signalized intersections): Free/uncoordinated operations. Using delay categories summarized in Exhibit 5­41, Exhibit 5­42 summarizes the change in delay that can be expected for heavy movements and DDI movements overall by applying the various strategies listed above. The results are based on a simulation­based experiment using calibrated base models of a DDI employing the three­critical­movement phasing scheme. The exhibit rows list the different strategies, while the columns list the six heavy­volume scenarios. The expected change in delay is expressed relative to base performance for the particular heavy­ volume scenario using a five­point scale (summarized in Exhibit 5­41). Blank cells indicate scenarios that were not specifically evaluated. These tables should be used to prioritize strategy testing but should not be used to exclude specific strategies for particular locations. More information about each strategy is available in the following sections. 5.4.4.1 Half Cycle Using half cycles at some intersections may provide for improved progression of off­peak traffic at the DDI by opening the green band more often. Additionally, this may help discourage red­light running at DDI intersections; because adjacent intersections are likely to require longer cycle lengths, a half cycle DDI can help reduce driver frustration due to long wait times at the crossover. Half cycling may also be useful in progressing both the through movement from the adjacent upstream signalized intersection and a heavy left­turn movement from the side street at that intersection. While this guidance only tested two cycles of equal length that summed to the overall corridor cycle length, it is also possible to run multiple cycle lengths at the DDI, provided the sum of those cycle lengths equals the cycle length of the overall corridor. 5.4.4.2 Signalized On-Ramp Left Turn Pedestrian facilities that utilize the outside of the DDI must cross a free­flow left turn onto the entrance ramp. For pedestrians, this movement could be signalized with a dedicated pedestrian Symbol Delay Category Associated Impact on Vehicle Operations Associated Percent Change in Delay ++ High Delay Increase Worse Conditions >12% Increase in Delay + Low Delay Increase Moderately Worse Conditions 6% –12% Increase in Delay o Minimal Delay Change Insignificant Change in Conditions <6% Increase in Delay and <6% Decrease in Delay - Low Delay Decrease Moderately Improved Conditions 6% –12% Decrease in Delay -- High Delay Decrease Improved Conditions >12% Decrease in Delay (Blank) Not Evaluated Not Evaluated Not Evaluated Exhibit 5-41. Delay categories.

Conceptual Operations 107 phase or a beaconing device such as a pedestrian hybrid beacon (PHB) or flashing yellow beacon. To avoid creating a left­turn queue between the crossovers, the pedestrian signal could be coor­ dinated with the DDI signal. While not specifically a corridor improvement strategy, pedestrian signals were included here to evaluate whether this pedestrian safety treatment would result in a significant negative impact on corridor operations. The results showed that the left turns can readily be signalized to enhance pedestrian safety without negatively impacting the operations. A practitioner should consider the impacts of shared versus exclusive lanes when evaluating this strategy. In a shared­lane scenario, the impact of queue spillback under heavy volumes could result in reduced throughput for the through movement. Strategy Low Volume Heavy Through Heavy Left Off Heavy Right Off Heavy Left On Heavy Through + Right He av y M ov em en t Al l M ov em en ts He av y M ov em en t Al l M ov em en ts He av y M ov em en t Al l M ov em en ts He av y M ov em en t Al l M ov em en ts He av y M ov em en t Al l M ov em en ts He av y M ov em en t Al l M ov em en ts Half Cycle ++ - ++ ++ o -- -- ++ ++ o - o Signalized On-Ramp Left Turn o o o o o o Dedicated Phase for Concurrent Off-Ramp Left and Right Turns o o o ++ ++ o RTOR Allowed at Off-Ramp o - o o -- -- - o Left-Turn-on-Red (LTOR) Allowed at Off-Ramp - -- o o -- -- o - LTOR & RTOR Allowed at Off- Ramp - -- o + o -- o -- Dynamic Overlap Phasing o o - o o o Alternate Side-Street Phases at Downstream Signal o o ++ + - o + o Lead/Lag Phasing for Outbound Lefts at Downstream Signal o o -- o + o -- - Eliminate Phases at Adjacent Intersection ++ + -- -- o ++ -- - Free/Uncoordinated ++ + ++ ++ ++ o ++ ++ ++ + ++ High Delay Increase + Low Delay Increase o Minimal Delay Change - Low Delay Decrease -- High Delay Decrease Exhibit 5-42. Change in delay expected with various strategies and traffic patterns (based on the three-critical-movement phasing scheme). Exhibit 5-42 is based on the three-critical-movement phasing scheme; different results would be expected if one of the other phasing schemes was chosen.

108 Diverging Diamond Interchange Informational Guide 5.4.4.3 Dedicated Phase for Concurrent Off-Ramp Left and Right Turns For continuously high demand at the exit ramp during peak hours, an additional phase can be added to the time­of­day plan, which serves the left and right turns from the exit ramp. This strategy uses a similar approach as was discussed in Section 5.3.4.3, but instead of a true “ramp flush,” it would use a phase to serve the ramp movements concurrently. 5.4.4.4 LTOR or RTOR Allowed at Off-Ramp Assuming proper sight distance is provided and drivers can identify the stream of traffic with which they are merging, RTOR or LTOR can be considered. If there are sufficient gaps in the cross­street traffic, the allowance of turning on red would increase the capacity of the movement. As with any interchange and closely­spaced intersection pair (i.e., not unique to DDIs), there is the potential for a weaving maneuver as vehicles turning off of the freeway quickly maneuver over multiple lanes of traffic to turn at the adjacent intersection. Where LTOR is not legal, the use of flashing arrows could be considered. 5.4.4.5 Dynamic Overlap Phasing Dynamic overlap phasing allows more efficient clearance of crossover through movements. Detection allows the clearance interval to be timed based on real­time vehicle information instead of estimated travel time. This can result in more time for the off­ramp movements while still providing clearance for the cross­street vehicles. 5.4.4.6 Alternate Side-Street Phases at Downstream Adjacent Signalized Intersection The phasing scheme at an intersection adjacent to a DDI could be adjusted to provide addi­ tional capacity for the cross­street approach while only allowing one or more side­street move­ ments every other cycle. This unusual phasing scheme could be used in a time­of­day plan when through traffic demand is high and when there is a great need for additional capacity to prevent queue spillback into the DDI. 5.4.4.7 Lead/Lag Phasing for Outbound Left Turns at Downstream Adjacent Signalized Intersection Lead/lag phasing allows signal timers flexibility to choose when to provide the left­turn phase in a specific ring so that the maximum bandwidth can be achieved for the coordinated move­ ment. Lead/lag phasing is being utilized much more frequently by agencies now that safety concerns with the “yellow trap” have been addressed through flashing yellow arrows and other innovations. In a DDI corridor, lead/lag phasing at the adjacent signalized intersection can help with progression of vehicle traffic to or from the DDI. 5.4.4.8 Eliminate Phases at Adjacent Signalized Intersection For upstream and downstream adjacent intersections, additional cross­street capacity could be achieved by eliminating phases. Eliminating phases is possible through a variety of alterna­ tive intersection forms such as the Restricted Crossing U­turn (RCUT) or the Median U­turn (MUT). With fewer phases, the available capacity at the adjacent intersection can be increased to levels similar to the very efficient DDI, thereby improving overall corridor operations. 5.4.4.9 Free/Uncoordinated In some cases, it may be more effective to allow the corridor to operate without coordina­ tion. For example, this may be beneficial during low­volume times of day. The delay results in

Conceptual Operations 109 Exhibit 5­42 suggest that this strategy is not appropriate for any of the high­volume scenarios because it results in significant increases in delay over the corresponding base case. However, for low­volume periods this strategy can more readily adapt to changing traffic patterns and has been applied successfully at several DDIs in the United States during off­peak periods. 5.5 Operational Analysis To support decisions regarding the choice and design of a DDI, there needs to be an appro­ priate level of traffic operational analysis corresponding to the stage of the project development process. This operational analysis should allow a practitioner to assess lane configurations, queues, and delay, with the ultimate goal of providing guidance to the designer. While most operational analysis focuses on motorized vehicles, final intersection configurations and associated signal timing should balance an agency’s priorities (e.g., multimodal users). The level of analysis should be consistent with available data, which could include: • Average daily traffic (ADT), • Future growth projections for the DDI and corridor (including traffic from planned or proposed developments adjacent to the interchange), • Speed (posted, design, or 85th percentile), • Weekday and weekend peak­hour turning­movement counts or demand estimates, • Weekday and weekend off­peak turning­movement counts or demand estimates, • Pedestrian volumes or demands, • Bicycle volumes or demands, • Estimates or counts of the proportion of the traffic stream composed of heavy vehicles, and • Origin­destination (O/D) demands at the DDI and adjacent signals. According to FHWA’s “Traffic Analysis Tools” web page, several tools are available to analyze traffic operations at intersections, including the following (6): • Planning­level analysis [such as critical­movement analysis (CMA) or Capacity Analysis for Planning of Junctions (CAP­X)], • HCM6 analysis, and • Microsimulation analysis. One major factor distinguishing these three types of analysis is the time required to complete the evaluation. HCM6 analysis may take several times as long as a planning­level analysis, and the time required for microsimulation is typically an order of magnitude greater than HCM6 analysis. In general: • Planning-level tools are useful for assessing the feasibility of a DDI and for selecting an initial number of lanes and phasing scheme. • An operational analysis using a deterministic method, such as the HCM6, is useful to perform a more detailed peak­hour performance analysis and to estimate performance measures like delay, travel time, and queue lengths (5). HCM6 analysis may provide insight on additional geometric design and signal timing details. • Microsimulation is useful for analyzing interactions between intersections and between modes to assess the overall performance in a multimodal corridor context. 5.5.1 Selecting the Appropriate Level of Analysis Selecting the appropriate level of analysis for a DDI evaluation depends on the available input data, the desired outputs, and the available resources to conduct the analysis. As such,

110 Diverging Diamond Interchange Informational Guide a DDI analysis does not differ fundamentally from any other interchange evaluation, but some DDI­specific considerations for tool selection are described below. 1. Pre-Screening the DDI as an interchange alternative. CAP-X, developed by FHWA, is a planning­level tool that can be used to quickly and directly compare the DDI to other (alternative) intersection and interchange forms (7). 2. Determining initial lane configuration and signal parameters. CMA, presented in NCHRP Report 812: Signal Timing Manual, 2nd Edition, allows for a quick evaluation of the required number of lanes at a DDI, estimation of basic signal timing parameters for the DDI, and screening for potential internal queuing issues based on minimal inputs (1). 3. Estimating interchange delay and Level of Service (LOS). The HCM6 DDI Method allows for estimation of interchange delays, queue lengths, and LOS using a deterministic procedure calibrated from DDI field observations in the United States (5). 4. Evaluating DDI corridor and multimodal performance. Microsimulation is ideally suited for evaluating DDIs in a corridor context, including interactions with adjacent intersections and different transportation modes. The selected microsimulation should allow sufficient flexibility and level of detail to replicate the DDI geometry and signal timing characteristics, and the analyst should consult the software user manual to assure that the tool is appropriate for modeling DDIs. A summary of the inputs, outputs, levels of effort, and limitations is provided for each analysis tool in Exhibit 5­43. 5.5.2 Planning-Level Analysis Two principal types of planning­level analysis apply to DDIs. CAP­X is a sketch­planning tool that can be used to compare a DDI to other intersection and interchange alternatives using volume­to­capacity ratios. CMA is a planning­level tool that is useful for comparing signal phasing options and verifying the lane configuration. Planning­level analyses can be performed by hand or software. 5.5.2.1 Advantages and Disadvantages The key advantage of planning­level­analysis approaches is that the methods can generally be applied quickly and with minimal resources. The resulting high­level assessment of DDI performance can be helpful when determining the initial feasibility of a DDI or to explore the necessary number of lanes. However, a key disadvantage of these methods is that the simpli­ fications and assumptions are not appropriate for every situation. As a general rule, results from a planning­level analysis can be used to determine the next step. For example, if a planning­level analysis shows a volume­to­capacity ratio of 1.3 for a DDI, it is unlikely that a more detailed operational analysis would show such a design to work satisfactorily. Similarly, if the planning­level analysis shows a peak­hour volume­to­capacity ratio of 0.3, it is likely that the number of lanes could be reduced in the design. For any planning­ level results that are close to a given volume­to­capacity threshold, a more detailed operational analysis should be performed. 5.5.2.2 CAP-X CAP­X, developed by FHWA, is a tool that can be used to evaluate different types of innova­ tive junction designs (eight intersections, five interchanges, and three roundabouts) using given peak­flow volumes (7). Exhibit 5­44 is a screen capture from the CAP­X spreadsheet (7). Note that CAP­X refers to the DDI as a double­crossover diamond, or DCD.

Conceptual Operations 111 The intersections and interchanges are evaluated using critical­lane­volume summation. Implemented in a spreadsheet workbook, CAP­X is a simple and cost­effective sketch­planning tool that can help users focus on the most appropriate intersection/interchange forms for a particular location, prior to conducting a more demanding traffic simulation. CAP­X is not sensitive to signal timing, and as such, cannot be used to compare and contrast different DDI phasing schemes. For an alternative planning method that is sensitive to phasing schemes, refer to Section 5.5.1. The inputs for the CAP­X methodology include turning­movement counts at both crossovers, heavy­vehicle percentages, and the number of lanes at each point where two movements cross (the two crossovers, the two off­ramp right­turn entry points from the freeway, and the two off­ramp left­turn entry points from the freeway). The method further allows for an estimation of impacts from a future­year growth rate. The outputs for a DDI are the approximate volume­to­ capacity ratio at each of the six crossing points. A practitioner should pay special attention to the capacity assumptions used in the spreadsheet; they may not be appropriate in every situation. 5.5.2.3 Critical-Movement Analysis CMA is a planning­level methodology that can be used to identify the critical movements at a DDI and approximate the required cycle length. The example in this section assumes both Analysis Tool Level of Analysis Required Inputs Available Outputs Level of Effort Limitations CAP-X Pre-screening the DDI as an interchange alternative ▪ Number of lanes ▪ Hourly TMCs1 ▪ V/C2 ratios ▪ Comparison to other designs Low ▪ No delay and LOS ▪ No queues ▪ No signal timing ▪ No corridor effects ▪ No multimodal ▪ No lane utilization Critical- movement analysis Determining initial lane configuration and signal parameters ▪ Number of lanes ▪ Hourly TMCs ▪ V/C ratios ▪ Cycle lengths ▪ Queue check Low ▪ No delay and LOS ▪ No corridor effects ▪ No multimodal HCM6 DDI method Estimating interchange delay and LOS ▪ Number of lanes ▪ Hourly TMCs ▪ Signal timing ▪ V/C ratios ▪ Delay and LOS ▪ Queues Moderate ▪ No corridor effects ▪ No multimodal ▪ No signal timing optimization guidance3 Micro- simulation Evaluating DDI corridor and multimodal performance ▪ Number of lanes ▪ Corridor O/D4 volumes ▪ Signal timing ▪ Corridor data ▪ Ped/bike data ▪ Delay and LOS ▪ Queues ▪ Corridor performance ▪ Multimodal High ▪ No signal timing optimization guidance5 1 TMCs = turning-movement counts 2 V/C = volume-to-capacity 3 While the HCM6 does not contain any optimization routines, commercial implementations of the methods may provide optimization capabilities. Practitioners should ensure that optimization functions accurately reflect DDI operations. 4 O/D = origin-destination 5 Some microsimulation tools offer built-in optimization while others work in exchange with macro-level tools to provide optimization capabilities. Practitioners should ensure that optimization functions accurately reflect DDI operations. Exhibit 5-43. Comparison of DDI analysis tools.

112 Diverging Diamond Interchange Informational Guide off­ramp movements are signalized; free­flow movements should be treated differently. For more information on applying CMA, refer to NCHRP Report 812: Signal Timing Manual, 2nd Edition (1). This section explains the application of CMA for a DDI using two examples. In the first example (illustrated in Exhibit 5­45, Exhibit 5­46, and Exhibit 5­49), there is one lane for the westbound and eastbound left­turn movements from the freeway. In the second example (illustrated in Exhibit 5­47, Exhibit 5­48, and Exhibit 5­50), there are two lanes for these move­ ments. Exhibit 5­45 and Exhibit 5­47 depict the volumes and assumed lane configuration for each example. Exhibit 5­46 and Exhibit 5­48 depict the identification of the critical volume for each phase, and Exhibit 5­49 and Exhibit 5­50 summarize the critical volume for the entire interchange. Once practitioners have calculated the critical volume, they can use it to approximate the required cycle length using the equation below for estimated cycle length based on critical volume. Exhibit 5­51 uses the equation below to illustrate the volumes that can be accommodated under different cycle lengths for an eight­phase intersection (1). The two DDI examples assume the two­critical­movement phasing scheme is being used (with fixed­time phases for clearance). Capacity will vary depending on the headway and phasing scheme that is selected, so a prac­ titioner should apply the equation using site­specific values. 3600 3600 Cycle Length Sec Lost Time Per Cycle Sec Max Number of Vehicles Per Hour Headway Sec( ) ( ) ( ) ( ) = × − × Note: This diagram does not reflect the actual lane configuration of the Intersection V / C 0.17 V / C Project Number: 12345 Location City, State Date January 1, 2014 < 1200 1200 - 1399 1400 - 1599 0.14 V / C 0.17 V / C 0.13 V / C 0.16 V / C 0.17 Double Crossover Diamond Interchange (E-W) Design and Results Project Name: Cap-X Sample Evaluation 221 275 206 255 270 274 Critical Lane Volume Sum VOLUME / CAPACITY RATIO: 0.17 1600 Exhibit 5-44. CAP-X planning-level tool screen capture. Note that volumes have generally been distributed evenly across the lanes, with the exception of the north- bound and southbound left-turn movements. Even distribution is used for example purposes only, and an analyst should determine an appropriate distribution based on travel patterns.

Conceptual Operations 113 Exhibit 5-45. CMA volumes. Exhibit 5-46. CMA analysis.

114 Diverging Diamond Interchange Informational Guide Exhibit 5-47. CMA volumes with revised lane configuration. Exhibit 5-48. CMA analysis with revised lane configuration.

Conceptual Operations 115 Exhibit 5-49. CMA calculations. Exhibit 5-50. CMA calculations with revised lane configuration. Cycle Length (Seconds) Number of Cycles Per Hour Lost Time Per Cycle (Seconds)1 Effective Green Time Per Cycle (Seconds) Number of Vehicles Per Cycle2 Maximum Number of Vehicles Per Hour2 60 60 20 40 16 933 70 51 20 50 19 1000 80 45 20 60 23 1050 90 40 20 70 27 1089 100 36 20 80 31 1120 110 33 20 90 35 1145 120 30 20 100 39 1167 1 This lost time assumes that the intersection is operating with eight phases (four in each ring) with 5 seconds of lost time per phase. The lost time will be less at an intersection with fewer phases. 2 The number of vehicles that can be accommodated under the various cycle lengths was calculated assuming a flow rate of 1,400 vehicles per hour which correlates to a headway of approximately 2.5 seconds per vehicle, which is generally conservative for urban/suburban environments. Source: NCHRP Report 812: Signal Timing Manual, 2nd Edition (1) Exhibit 5-51. Estimated cycle lengths based on critical volume (eight-phase intersection).

116 Diverging Diamond Interchange Informational Guide Using Exhibit 5­51, a practitioner can identify that the critical volume in the first example (1,400 vehicles) requires a cycle length greater than 120 seconds. By adding a second lane for the off­ramp left­turn movements, the critical volume can be reduced in the second example (1,000 vehicles), requiring a shorter cycle length of approximately 70 seconds. Note that these cycle lengths depend heavily on the lost time and vehicle headways assumed at the DDI. The second step of a CMA for a DDI should be assessing the potential queues between the crossovers, using the equation below for vehicle queue length. Assuming that the lane configura­ tion from the second example is chosen (i.e., two lanes for the off­ramp left­turn movements), a cycle length of 70 seconds should be able to accommodate demand. Vehicle Queue Feet Critical Queued Volume Number of Cycles Per Hour Vehicle Length Feet( ) ( )=     With the two­critical­movement phasing scheme, there is the potential for the off­ramp left­turn movements to queue at the downstream crossover during Phases 4 and 8. In this example, the westbound left­turn movement from the off­ramp is heavier (500 vehicles), so it will dictate the required queue distance. With a cycle length of 70 seconds, there are 51 cycles per hour. Assuming an even distribution of traffic throughout the peak hour, there would be approximately 10 vehicles queued every cycle (= 500 vehicles/51 cycles per hour). Assuming a vehicle length of 25 feet, the crossovers should be at least 250 feet apart to mitigate queues. 5.5.3 HCM6 DDI Method HCM6 analysis tools are deterministic (similar to planning­level­analysis tools) but provide more detailed performance measures (i.e., delay, travel time, and queue lengths) at a lane­group level, as opposed to an overall intersection level. HCM6 analysis is performed using software, but individual calculations can be checked by hand. An HCM procedure specifically for DDIs is included in the Highway Capacity Manual 6th edition as part of Chapter 23: Ramp Terminals and Alternative Intersections (5). Because the chapter contains analytical methods for evaluating several intersection and interchange forms, the chapter and methodology are well­suited for a direct comparison of the DDI with other interchange or intersection forms. 5.5.3.1 Advantages and Disadvantages One of the advantages of this operational­level analysis is that it balances detail with reason­ able data input needs and analysis resource requirements. The HCM6 method provides more detailed output (i.e., delay, travel time, and queue estimates) than the planning­level methods, while allowing for more customization and consideration of geometric variability and signal timing details. At the same time, its methods are typically applied more quickly than a resource­ intensive simulation analysis. The deterministic analysis framework of the HCM6 methodology offers consistency in performance estimation across practitioners and interchange options. While the HCM6 has limitations, it does provide the consistency that agencies need for evaluating different alterna­ tives. The HCM6 DDI method includes several specific algorithms to reflect the operational characteristics of DDIs, including: • Estimation of saturation flow rates specific to DDIs. • Prediction of DDI lane utilization.

Conceptual Operations 117 • Calculation of internal DDI queue lengths. • Analysis of unsignalized turning movements at DDIs. • Consideration of unique attributes of DDI signal timing in the performance estimation, such as demand starvation and other adjustments to the effective green time. Disadvantages of the HCM6 analysis of DDIs include a limited scope of applicable geometry and a lack of focus on network and system effects, including the interaction of the DDI with the freeway facility it serves. Other operational characteristics of a DDI not handled by the existing HCM6 methodology and potentially requiring simulation­based analysis include: • Queuing on the links between the two crossover signals. • Demand starvation at signalized approaches leaving the DDI. • Queue blockage of on­ramp left­turn movements in shared lanes. • Impact of reverse curves on speed patterns and progression. • Estimation of pedestrian or bicycle LOS. • The effect of pedestrian and bicycle activity on vehicles. • The interaction between freeway and arterial traffic if queues from one facility impact operations on the other. Rather than replicating details of the methodology itself here, the reader is referred to the HCM6 for additional information (5). The method can be integrated with the HCM6 method on urban street facilities, which allows the HCM6 to evaluate a DDI in a corridor context. 5.5.4 Microsimulation Microsimulation tools employ a series of algorithms for car following, lane changing, and other parameters to model the movements and interaction of individual vehicles on a sub­ second interval basis. Most simulation algorithms are stochastic in nature, meaning that they include one or more random variables and distribution of variables, rather than a fixed deter­ ministic input (e.g., vehicle speed). For the evaluation of DDIs, simulation tools have been the primary analysis tool, as many tools are not able to directly account for the unique geometry and signal timing of this interchange form. While other planning and operational tools have become available and can be highly useful in alternative selection and design refinement, simulation remains inherently suited for DDI analysis. Microsimulation analysis is performed exclusively using software. A variety of simulation tools are available to model and evaluate DDIs. All microsimulation tools vary in user interface and available features. Among the more critical features that are required to accurately model a DDI is the ability to replicate the crossover geometry and accurately code the DDI signal timing sequence. The analyst should further review the list of calibration factors and validation parameters described in Section 5.5.4.2 to assure that the selected tool can adequately provide these. 5.5.4.1 Advantages and Disadvantages Simulation tools allow for flexible customization and configuration of geometry, signal timing, and other operational parameters. This allows for a direct estimation of DDI performance, rather than approximating certain effects through equations that may have been derived based on only a few sites. Using simulation, a DDI can be evaluated as part of a broader network of intersections, including the interaction between the cross street and the freeway. Many simu­ lation tools further allow the modeling and evaluation of different modes of transportation (e.g., pedestrian, bicycles, and transit) and their interaction with vehicular traffic. In addition, simulation provides visualization of traffic patterns and street geometry, which can be an invalu­ able asset for communicating the DDI to a nontechnical stakeholder audience.

118 Diverging Diamond Interchange Informational Guide The greatest disadvantage of simulation is the increased resource requirements, as every DDI model needs to be built and configured. Specifically, proper calibration and validation of the simulation model is needed to assure that the predicted operations adequately reflect conditions at the DDI. For a proposed site, guidance in the literature can help with the calibration process, some of which is also summarized below (8). In addition, not all microsimulation tools have the ability to optimize traffic signal timing, which is critical to obtain good signal operations and coordination for the DDI and surrounding intersections. As such, simulation may be used in conjunction with a signal timing and optimi­ zation tool to obtain signal timing parameters for the network. That optimization tool needs to be able to accommodate the DDI­specific geometry and signal configuration, but otherwise, the practitioner can follow the normal timing and optimization process used for other inter­ sections and interchanges. The practitioner needs to understand the many unique operational attributes of a DDI (e.g., saturation flow rate, speed profiles, and lost time) as well as how to replicate those in simulation. This lack of consistency in output can be an important limitation of simulation (i.e., different results from different practitioners), especially for DDIs with estimated per­ formance close to a defined threshold. In this case, a deterministic analysis method could be used in combination with simulation to further inform the decision­making process. 5.5.4.2 Calibration Factors Practitioners use calibration to adjust models prior to construction so that the predicted out­ come is as accurate as possible. Each simulation tool has many calibration factors, ranging from demand inputs to speed settings to signal timing parameters. The discussion that follows is not intended to be an exhaustive or complete list for simulation calibration, but highlights specific calibration settings that have proven necessary to adequately replicate DDI operations. All typical calibration steps should be followed. The key calibration factors for DDIs include: • O/D demands at the DDI and adjacent signals that are based on field data, • Look­back distances from route decision points to control lane positioning, • Field­measured free­flow speeds through the DDI, as well as geometrically­constrained free­flow speeds at the crossover and for turning movements, and • DDI­specific phasing schemes as obtained from field controller settings or design plans at the DDI and adjacent signals. 5.5.4.2.1 Origin-Destination Demands. Interchange O/D route percentages are no different at a DDI than at a conventional interchange but are important for accurately representing operations at a DDI. O/D routes should be drawn through the entire interchange and adjacent signalized intersections. 5.5.4.2.2 Look-Back Distances. The look­back distance is the distance upstream from a diverge point at which simulated vehicles are affected and initiate any necessary lane changes. In field observations of DDIs, drivers were observed to pre­position themselves well in advance of the DDI for downstream turning movements (2). This phenomenon was especially pro­ nounced for left­turning movements from the cross street to the freeway, and lane utilization was impacted at the inbound DDI crossover. Consequently, the look­back distance for these movements should be specified in a way that it extends through the inbound DDI crossover (as shown in Exhibit 5­52). 5.5.4.2.3 Free-Flow Speeds. As shown earlier in this chapter, speeds at DDI crossovers were observed to be below the free­flow speeds on tangent sections of the cross street. It is therefore recommended that the analyst use speed­reduction zones to control free­flow speeds at

Conceptual Operations 119 Exhibit 5-52. Simulation look-back distance. crossovers and turns. The speed distributions should be modeled as normal distributions with the mean and standard deviation estimated from field data or adapted from the discussion in Section 5.4.2. Exhibit 5­53 shows the placement of reduced­speed areas for modeling slower speeds at turns and crossovers. 5.5.4.2.4 Phasing Schemes. To accurately model signalized control of a DDI, a practitioner needs to have a clear understanding of the phasing scheme so he or she can confirm whether the scheme can be represented accurately in the chosen simulator. The selected tool should employ signal control logic that is flexible enough to allow modeling of all three DDI phasing schemes (and their variations) introduced in Section 5.3. Because the phasing schemes can be implemented in either one or two controllers, the simulator should also be able to model either configuration. 5.5.4.3 Validation Parameters Validation allows practitioners to refine models to reflect existing conditions. This section explains targets that can be used to validate models against conditions at constructed DDIs. Similar to the prior section on calibration, the discussion that follows is not intended to be an exhaustive or complete list for validation but rather to highlight specific measures that have proven useful for quantifying and validating DDI operations. Other validation steps and factors Exhibit 5-53. Simulation speed settings.

120 Diverging Diamond Interchange Informational Guide should be followed as well. Three key validation parameters recommended in the literature for accurately modeling DDIs include (8): 1. Interchange travel times for the through movements (between the crossovers) and left­turn movements (to and from the freeway). 2. Route travel times for through movements (through the DDI and adjacent signals) and left­turn movements (to and from the freeway through the adjacent signals). 3. Comparison of average and 95th percentile queue lengths (estimated from maximum queue lengths on a per cycle basis). 5.5.4.3.1 Travel Times. The interchange travel time includes the two DDI crossover signals and any queues immediately upstream of the DDI. The route travel time segments include, at a minimum, the adjacent signalized intersections upstream and downstream of the DDI. For left­turn routes, the travel time segments start or end at the freeway exit ramp or entrance ramp, respectively. Travel time data can be collected in a variety of ways (e.g., floating­car technique). 5.5.4.3.2 Queue Lengths. For queue measurements, cycle­by­cycle queues can be observed through manual observations in the field on a per lane basis. Detailed operational study protocols are documented in Field Evaluation of Double Crossover Diamond Interchanges (2). 5.6 References 1. Urbanik, T., A. Tanaka, B. Lozner, E. Lindstrom, K. Lee, S. Quayle, S. Beaird, et al. NCHRP Report 812: Signal Timing Manual, 2nd ed., Transportation Research Board, Washington, D.C., 2015. 2. Cunningham, C., B. Schroeder, J. Hummer, C. Vaughan, C. Yeom, K. Salamati, D. Findley, J. Chang, N. Rouphail, S. Bharadwaj, C. Jagadish, K. Hovey, and M. Corwin. Field Evaluation of Double Crossover Diamond Interchanges. Contractor’s Draft Submittal. FHWA, Project No. DTFH61­10­C­00029, 2014. 3. Yeom, C., B. J. Schroeder, C. Cunningham, K. Salamati, and N. M. Rouphail. Lane Utilization Model Devel­ opment for Diverging Diamond Interchanges. Presented at the 95th Annual Meeting of the Transportation Research Board, Washington, D.C., 2016. 4. Yeom, C., B. Schroeder, C. Cunningham, C. Vaughan, N. Rouphail, and J. Hummer. Lane Utilization at Two­Lane Arterial Approaches to Double Crossover Diamond Interchanges. Transportation Research Record: Journal of the Transportation Research Board, No. 2461, 2014, pp. 103–112. 5. Highway Capacity Manual: A Guide for Multimodal Mobility Analysis, 6th ed. Transportation Research Board, Washington, D.C., 2016. 6. FHWA. “Traffic Analysis Tools” web page. http://ops.fhwa.dot.gov/trafficanalysistools. Accessed June 22, 2016. 7. FHWA. “Capacity Analysis for Planning of Junctions (Cap­X) Tool” web page. https://www.fhwa.dot.gov/ software/research/operations/cap­x/. Accessed June 22, 2016. 8. Schroeder, B., K. Salamati, and J. Hummer. Calibration and Field Validation of Four Double­Crossover Diamond Interchanges in VISSIM Microsimulation. Transportation Research Record: Journal of the Transpor- tation Research Board, No. 2404, 2014, pp. 49–58.