National Academies Press: OpenBook
« Previous: Chapter Four - Cross-Section Elements
Page 35
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 35
Page 36
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 36
Page 37
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 37
Page 38
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 38
Page 39
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 39
Page 40
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 40
Page 41
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 41
Page 42
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 42
Page 43
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 43
Page 44
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 44
Page 45
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 45
Page 46
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 46
Page 47
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 47
Page 48
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 48
Page 49
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 49
Page 50
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 50
Page 51
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 51
Page 52
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 52
Page 53
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 53
Page 54
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 54
Page 55
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 55
Page 56
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 56
Page 57
Suggested Citation:"Chapter Five - Intersections." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 57

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.

36 Overview Design of and warrants for auxiliary lanes were common topics of interest during the past decade. Researchers also revisited effects of and treatments for skewed approaches, channelization, and other intersection configuration elements. Consideration of bicycles and pedestrians at intersections was promoted through a variety of initiatives, particularly related to pedestrians with disabilities. The subject that captured the attention of many, however, was modern roundabouts; the design, installation, and operation of these alternatives to signalization were the subject of much research, and lead to the publication of two FHWA Guides during this time. In addition, other innovative intersection designs appeared in research and on highway networks; many of these new inter- sections are designed with the purpose of improving capacity by changing the manner in which left-turn movements are accommodated. Finally, access management near intersections also garnered a great deal of interest, as researchers looked for ways to minimize impacts of adjacent driveways and side streets. intersectiOn cOnfiguratiOn Kindler et al. (2004) described the development of an expert system for diagnostic review of at-grade intersections on rural two-lane highways. This system, the Intersection Diag- nostic Review Module (IDRM), was developed as a compo- nent of the Interactive Highway Safety Design Model to aid designers in assessing the safety consequences of geometric design decisions, particularly for combinations of geomet- ric features. IDRM was developed to allow such problems to be identified and evaluated in an automated and orga- nized fashion. IDRM identifies concerns by “using models of the criticality of specific geometric design situations. These include existing geometric design models—such as sight distance models—as well as newly developed models. IDRM uses 21 specific models to address 15 high-priority issues related to [the] intersection as a whole and [to indi- vidual] approach” legs. “IDRM makes no attempt to select a particular treatment as appropriate to the intersection. After further investigation, the IDRM user may select a particular treatment as appropriate [on the basis of the] available evi- dence and engineering judgment, or the user may conclude that no treatment is necessary and that the project should be built as designed.” alignment FHWA’s Signalized Intersections: Informational Guide (Rodegerdts et al. 2004) states that “the approach to a sig- nalized intersection should promote awareness of an inter- section by providing the required stopping sight distance in advance of the intersection.” The document recommends the following guidelines to meet drivers’ and cyclists’ expecta- tions as they approach intersections: • “Avoid approach grades to an intersection of greater than 6%. On higher design speed facilities (50 mph and greater), a maximum grade of 3% should be considered. • Avoid locating intersections along a horizontal curve of the intersecting road. • Strive for an intersection platform (including sidewalks) with cross slope not exceeding 2%, as needed for accessibility.” Approach curvature is a geometric design treatment that can be used at high-speed intersection approaches to force a reduction in vehicle speed through the introduction of hori- zontal deflection, as described in NCHRP Report 613 (Ray et al. 2008). As shown in Figure 9, approach curvature con- sists of successive curves with progressively smaller radii. Research and applications of approach curvature previously focused on roundabouts. However, the report states that this geometric design treatment has potential to be applied to conventional intersections as well. The use of approach curvature at downhill approaches was discouraged. The report authors’ experience with approach curvature suggested that this geometric treatment can be used in conjunction with reduced speed limit signs or advi- sory speed signs. The length and curve geometry can be determined from the upstream segment operating speed and the target speed and appropriate design vehicle for the intersection. effect Of skew Son et al. (2002) developed a methodology for calculating sight distance available to drivers at skewed unsignalized intersec- tions. The methodology considered that the sight distance may vary depending on the driving positions of the drivers and the different lines of sight given to drivers by different chapter five intersectiOns

37 types of vehicles with unique sight-line obstructions. They derived equations and nomographs for calculating available sight distance that included the influence of factors such as geometry (intersection angle, lane width, shoulder width, position of stop line), vehicle dimensions, and the driver’s field of view. They concluded that findings from their research provided evidence that a skew angle greater than 20 degrees should not be used in design when the design vehicle is a large vehicle or semitrailer, because available sight distances were less than the required stopping sight dis- tance, even with a low value of design speed for intersection angles less than 70 degrees. The Highway Design Handbook for Older Drivers and Pedestrians (Staplin et al. 2002) recommends establishing 15 degrees as a minimum skew angle as a practice to accom- modate age-related performance deficits at intersections where right-of-way is restricted; “at skewed intersections where the approach leg to the left intersects the driver’s approach leg at an angle of less than 75 degrees, the prohibition of right turn on red (RTOR) is recommended.” The Handbook cites multiple studies documenting restricted neck move- ment in older drivers, making detection of and judgments about potential conflicting vehicles on crossing roadways much more difficult. auxiliary lane Design Harwood et al. (2002) conducted a study to investigate the safety effectiveness of left- and right-turn lane treatments. The research team collected geometric design, traffic con- trol, traffic volume, and traffic crash data at 280 improved sites in eight states and at 300 similar intersections that were not improved during the study period. The types of improve- ment projects evaluated included installation of added left- turn lanes, installation of added right-turn lanes, installation of added left- and right-turn lanes as part of the same proj- ect, and extension of the length of existing left- or right-turn lanes. Based on the results of the analyses, they concluded the following: • Adding left-turn lanes is effective in improving safety at signalized and unsignalized intersections. Installing a single left-turn lane on a major-road approach would be expected to reduce total intersection accidents at rural unsignalized intersections by 28% for four-leg intersections and by 44% for three-leg intersections, with corresponding reductions of 27% and 33% at urban unsignalized intersections. At four-leg urban signal- ized intersections, installation of a left-turn lane on one approach would be expected to reduce accidents by 10%, and installation on both major-road approaches would be expected to increase, but not quite double, the resulting effectiveness measures for total intersection accidents. • Positive results can also be expected for right-turn lanes, with reductions in total intersection accidents of 14% at rural unsignalized locations and 4% at urban signalized locations for installations on single approaches. “Instal- lation of right-turn lanes on both major-road approaches to four-leg intersections would be expected to increase, but not quite double, the resulting effectiveness mea- sures for total intersection accidents.” • “In general, turn-lane improvements at rural intersec- tions resulted in larger percentage reductions in acci- dent frequency than comparable improvements at urban intersections.” • “Overall, there [was] no indication that any type of turn-lane improvement is either more or less effective for different accident severity levels.” FHWA’s Highway Design Handbook for Older Drivers and Pedestrians (Staplin et al. 2002) states that “two fac- tors can compromise the ability of older drivers to remain within the boundaries of their assigned lanes during a left turn. One factor is the diminishing ability to share atten- tion (i.e., to assimilate and concurrently process multiple sources of information from the driving environment). The other factor involves the ability to turn the steering wheel sharply enough, given the speed at which they are traveling, to remain within the boundaries of their lanes.” Data sources cited by the Handbook’s authors indicated that a 12-ft lane width provides the most reasonable tradeoff between the FIGURE 9 Example of approach curvature (Robinson et al. 2000).

38 need to accommodate older drivers, as well as larger turning vehicles, without penalizing the older pedestrian in terms of exaggerated crossing distance. The Handbook’s correspond- ing recommendation was for a minimum receiving lane width of 12 ft accompanied, wherever practical, by a shoul- der of 4 ft minimum width. The Handbook further recommended that, for new or reconstructed facilities, unrestricted sight distance, achieved through positive offset of opposing left-turn lanes, be pro- vided whenever possible. This recommendation was made in anticipation of providing “a margin of safety for older driv- ers who, as a group, do not position themselves within the intersection before initiating a left turn.” Where the provi- sion of unrestricted sight distance is not feasible, positive left-turn lane offsets were recommended to achieve the mini- mum required sight distances appropriate for major roadway design speed and type of opposing vehicle. Long (2002) reviewed the characteristics of intervehicle spacing for the purpose of auxiliary lane design. He concluded that the value of 25 ft per vehicle used by the CORSIM mod- eling software was a severe underestimation for determining queue lengths, as was the 3-ft distance between vehicles. He developed new models for estimating average queue lengths and maximum lengths at a given probability; models based on an intervehicle spacing of 12 ft, a passenger car length of 15 ft, a 65-ft length for combination trucks, and 30 ft for other vehicles. Kikuchi et al. (2005) developed a method for estimating the needed length of dual left-turn lanes (DLTL). Their pro- cedure first surveyed how drivers choose a lane of the DLTL in the real world and analyzed the relationship between lane use and the volume of left-turn vehicles. Second, the pro- cedure calculated the probability that all arriving left-turn vehicles during the red phase could enter the left-turn lanes (i.e., no queue spillback of vehicles from the DLTL and no blockage of the DLTL by the queue of through vehicles). This probability was presented as a function of the length of the DLTL and the arrival rates of left-turn and through vehicles. The adequate lane length was derived such that the probability of the vehicles entering the DLTL is greater than a threshold value. Third, the recommended adequate length was expressed in number of vehicles, then converted to the actual distance required based on the vehicle mix and prefer- ence between the two lanes. Resulting recommended lengths were presented as a function of left-turn and through volumes for practical application. Lee et al. (2005) developed models to predict lane utiliza- tion factors for six types of intersections with downstream lane drops and to assess how low lane utilization affects the observed intersection capacity and level of service. They col- lected traffic and signal data at 47 sites in North Carolina. On the basis of 15 candidate factors, multiple regression mod- els were developed for predicting the lane utilization factor. They compared field-measured delays with delays estimated by the HCM with the use of regression models for lane uti- lization. They stated that even with the new models for lane utilization, the HCM consistently overestimated delay for all types of lane-drop intersections with low lane utilization and suggested that a reassessment of the effect of lane utilization on capacity may be in order. The study also found that the downstream lane length and traffic intensity positively corre- lated with the lane utilization factor, existence of a TWLTL or midblock left-turn bay increased the lane utilization fac- tor, lane drops resulting from lane usage change had more equal lane volume distribution than the midblock taper lane drop, and that some geometric variables at the approach may also influence lane utilization. Fitzpatrick et al. (2006) conducted a study to determine variables that affected the speeds of free-flow turning vehicles in an exclusive right-turn lane and explore the safety expe- rience of different right-turn lane designs. Their evaluations found that the variables affecting the turning speed at an exclu- sive right-turn lane included the type of channelization pres- ent (either lane line or raised island), lane length, and corner radius. Variables that affected the turning speed at an exclu- sive right-turn lane with island design included (1) radius, lane length, and island size at the beginning of the turn, and (2) corner radius, lane length, and turning-roadway width near the middle of the turn. The authors compared this with previous research treatments that had the highest number of crashes were right-turn lanes with raised islands. In their anal- ysis, they found this type of intersection had the second high- est number of crashes of the treatments evaluated in this study. In both studies, the “shared through with right lane combina- tion” had the lowest number of crashes. They recommended that these findings be verified through use of a larger, more comprehensive study that includes right-turning volume. NCHRP Project 3-72 was tasked with developing design guidance related to right-turn lanes on urban and suburban arterials. The research team from Midwest Research Insti- tute discussed results from their research with respect to right-turn deceleration lanes (Potts et al. 2007a). They con- ducted a computer simulation study of motor vehicles and pedestrians at right-turn lanes to determine their operational effects. They also performed a benefit-cost analysis of right-turn lanes that considered both operational and safety effects. The researchers determined that right-turn maneuvers from a two-lane arterial at an unsignalized intersection or driveway can delay through traffic by 0 to 6 s per through vehicle where no right-turn lane was present. Delays to through traffic owing to right turns in the same situation on a four-lane arterial were substantially lower, in the range from 0 to 1 s per through vehicle. They concluded that pedestrians at unsignalized intersections or driveways can have a substantial impact on delay to through vehicles owing to slowing of right-turning vehicles yielding to pedes- trians, but provision of a right-turn lane could reduce pedestrian-related delays to through traffic by as much as

39 6 s per through vehicle, depending on pedestrian volume. Results from the project’s economic analysis procedure developed a method to identify where installation of right- turn lanes at unsignalized intersections and major drive- ways would be cost-effective, indicating combinations of through-traffic volumes and right-turn volumes for which provision of a right-turn lane would be recommended. The research team stated that their economic analysis proce- dure can be applied by highway agencies using site-specific values for ADTs, turning volumes, accident frequency, and construction cost for any specific location (or group of similar locations) of interest. Kikuchi et al. (2007) examined the lengths of turn lanes when a single lane approached a signalized intersection and was divided into three lanes: left-turn, through, and right- turn. Their objective was to determine the appropriate length of each turn lane. From analysis of the vehicle queue pat- tern at the entrance to the turn lanes, they developed a set of formulas to compute the probabilities of the occurrence of turn-lane overflow and turn-lane blockage. The recom- mended lane lengths were calculated so that the probabili- ties that a lane did not overflow and that the entrance of the lane was not blocked were greater than a threshold value of 0.95. Recommended turn-lane lengths, presented in a series of tables, were found to be shorter than those recommended by AASHTO. In a subsequent study, Kikuchi and Kronprasert (2008) developed analytical and computational processes for deter- mining the length of the right-turn lane at a signalized inter- section. They examined the factors that influenced length, reviewed available literature and practices, derived recom- mended lengths analytically, and developed a set of tables of recommended lane lengths as a function of approach volumes (right-turn, through-traffic, and cross-traffic volumes) and signal timing. Their analysis compared conditions when right-turn-on-red (RTOR) was not permitted and when it was permitted. Based on achieving desired probabilities of turn-lane overflow and turn-lane blockage, they calcu- lated recommended lane lengths based on the number of vehicle spaces and described a procedure to convert that number to actual distance. They compared their guidelines that account for arrival rates of both right-turn and through vehicles with guidelines that only considered right-turn vehicles; as a result, they concluded their proposed lane lengths were different than those in existing guidelines. Their recommended lengths for RTOR conditions were somewhat shorter than non-RTOR conditions when the right- turn arrival rate was greater than the arrival rate for through vehicles. FHWA’s Highway Design Handbook for Older Drivers and Pedestrians (Staplin et al. 2002) recommends “raised channelization with sloping curbed medians rather than channelization accomplished through the use of pavement markings, for the following operating conditions: • Left- and right-turn lane treatments at intersections on all roadways with operating speeds of less than 40 mph. • Right-turn treatments on roadways with operating speeds equal to or greater than 40 mph.” Where raised channelization is implemented at intersections, they also recommended that median and island curb sides and curb horizontal surfaces be treated with retroreflectorized markings and be maintained at a minimum luminance contrast level of 2.0 with overhead lighting or 3.0 without overhead lighting. intersectiOn sight Distance Where determinations of intersection sight distance (ISD) requirements for any intersection maneuver that is performed by a driver on either a major or a minor road incorporate a PRT component, the FHWA Highway Design Handbook for Older Drivers and Pedestrians (Staplin et al. 2002) rec- ommends that a PRT value of no less than 2.5 s be used to accommodate the slower decision times of older drivers. It also recommends that “where determinations of intersection sight-distance requirements for a left-turn maneuver from a major roadway by a stopped passenger car are based on a gap model, a gap of no less than 8.0 s, plus 0.5 s for each addi- tional lane crossed by the turning driver, be used to accom- modate the slower decision times of older drivers.” Yan and Radwan (2005) conducted research to develop sight distance geometric models for unprotected left-turning vehicles from the major road to the minor road at signalized intersections; they also sought to evaluate sight improvement effects of two offset methods and analyze the relationship between available sight distance and selected geometric param- eters. According to their conclusions, sight distance problems could occur for passenger cars on traditional left-turn lane designs with 14- to 18-ft medians at high design speeds. Using sensitivity analyses, they also developed equations showing a relationship between sight distance and offset value for parallel left-turn lanes and between sight distance and taper angle for taper lanes. Left-turn lane length was also cited as an important variable that affects sight distance. Easa and Ali (2006) developed an extension of a previous ISD model to consider sight distance for stop-control inter- sections on three-dimensional alignments. Although the pre- vious model accounted for obstructions inside the horizontal curve and for intersections and major-road vehicles (objects) on the curve, their model was expanded by (1) allowing the object to be anywhere on the horizontal curve or tangent, (2) allowing the horizontal and vertical curves to overlap par- tially, and (3) considering the case in which the obstruction lies outside the horizontal curve. The obstruction location was for- mulated through use of a simple variable that takes the value of +1 or -1 for an obstruction, respectively, inside or outside the horizontal curve. They presented design aids for the required minimum lateral clearances (from the minor and major roads)

40 for different radii of horizontal curve and major-road design speeds. They noted that their model considers the vertical obstruction caused by the road surface on crest vertical curves, and recommended further research to explore the case of a sag vertical curve, where the sightline may be obstructed by an overpass. mODern rOunDabOuts The increase in the use of modern roundabouts in the United States continued at a high pace during the decade from 2000 to 2010. The need to know more about design, operations, safety, and other aspects of roundabouts in this country prompted a number of research projects. Findings from those projects will be summarized in this section of the report. This section contains subtopics that overlap with headings found elsewhere in this chapter (e.g., design speed and alignment), but the roundabout-specific nature of the information made it appropriate to include here, rather than be distributed through- out other parts of the report. NCHRP Synthesis 299 had little content on roundabouts, because, to that point, relatively little research had been conducted on them in the United States. A series of projects during the decade led to the publication of two FHWA Informational Guides containing recommendations and guidelines for all aspects of roundabout design. As such, a great deal of content was generated on the subject, a sample of which is presented in this section. Many projects were regional or local in nature, however, and specific research reports on many of those projects were not published in a forum that was readily available for this synthesis. How- ever, the FHWA Informational Guides summarized much of the existing information and compiled them into the form of nationally distributed research reports as well as guidelines suitable for practitioners. Given the importance of those two Guides, a sizeable portion of the research highlighted here is either primarily or secondarily sourced to those two documents. general Principles The authors of NCHRP 672: Roundabouts: An Informational Guide, Second Edition (Rodegerdts et al. 2010) offered a set of overarching principles to guide the development of designs for all roundabouts. They stated that achieving these principles be the goal of any roundabout design: • “Provide slow entry speeds and consistent speeds through the roundabout by using deflection. • Provide the appropriate number of lanes and lane assign- ment to achieve adequate capacity, lane volume balance, and lane continuity. • Provide smooth channelization that is intuitive to drivers and results in vehicles naturally using the intended lanes. • Provide adequate accommodation for the design vehicles. • Design to meet the needs of pedestrians and cyclists. • Provide appropriate sight distance and visibility for driver recognition of the intersection and conflicting users.” Design speed The authors of the first edition of FHWA’s Roundabouts: An Informational Guide (Robinson et al. 2000) stated that design speed of a roundabout is determined from the small- est radius along the fastest allowable path. In their observa- tions, the smallest radius usually occurred on the path of the circulatory roadway as the vehicle curved to the left around the central island. However, they stated it was “important when designing the roundabout geometry that the radius of the entry path (i.e., as the vehicle curves to the right through entry geometry) not be significantly larger than the circu- latory path radius.” Recommended maximum entry design speeds from the Guide are shown in Table 17. The design process described in the Guide states that a vehicle is assumed to be 2 m (6 ft) wide and to maintain a minimum clearance of 0.5 m (2 ft) from a roadway centerline or concrete curb and flush with a painted edge line. Thus the cen- terline of the vehicle path is drawn with the following distances to the particular geometric features: • 1.5 m (5 ft) from a concrete curb, • 1.5 m (5 ft) from a roadway centerline, and • 1.0 m (3 ft) from a painted edge line. Their desirable radius relationship was that the entry path radius was less than the circulatory path radius, which was less than the exit path radius, ensuring that speeds will be reduced to their lowest level at the roundabout entry. The design speed review process also included the evaluation of the left-turn path radius and the right-turn path radius for speeds consistent with the other three radii. Selection of an appropriate design vehicle, as defined in the Green Book, would help to define the necessary radii for a given design speed. The second edition of Roundabouts: An Informational Guide (Rodegerdts et al. 2010) recommends maximum enter- ing design speeds based on a theoretical fastest path of 20 to 25 mph for single-lane roundabouts; 25 to 30 mph is rec- ommended for multilane roundabouts, based on a theoretical fastest path assuming vehicles ignore all lane lines. TABLE 17 RECOMMENDED MAxIMUM ENTRY DESIGN SPEEDS Site Category Recommended Maximum Entry Design Speed, km/h (mph) Mini-roundabout 25 (15) Urban Compact 25 (15) Urban Single Lane 35 (20) Urban Double Lane 40 (25) Rural Single Lane 40 (25) Rural Double Lane 50 (30) Source: Robinson et al. 2000.

41 alignment With regard to the alignment of roundabout approaches, the first FHWA Guide (Robinson et al. 2000) states that, in general, the roundabout is optimally located when the center- lines of all approach legs pass through the center of the inscribed circle. This location usually allows the geometry to be adequately designed so that vehicles will maintain slow speeds through both the entries and the exits. The radial alignment also makes the cen- tral island more conspicuous to approaching drivers. If it is not possible to align the legs through the center point, a slight offset to the left (i.e., the centerline passes to the left of the roundabout’s center point) is acceptable. It is almost never acceptable for an approach alignment to be offset to the right of the roundabout’s center point. This alignment brings the approach in at a more tangential angle and reduces the opportunity to provide sufficient entry curvature. Examples of all three alignments are shown in Figure 10. lane arrangement NCHRP Report 672 (Rodegerdts et al. 2010) provides a meth- odology for conducting an operational analysis of a round- about, one outcome of which is to determine the required number of entry lanes to serve each approach. The report’s authors advise that, for multilane roundabouts, care must be taken to ensure that the design also provides the appropri- ate number of lanes within the circulatory roadway and on each exit to ensure lane continuity. The primary caution with multilane roundabouts is path overlap, which occurs when the natural path through the roundabout of one traffic stream overlaps the path of another. If the natural path of one lane interferes or overlaps with the natural path of the adjacent lane, the roundabout is not as likely to operate as safely or efficiently as possible. The report advises that a good entry design aligns vehicles into the appropriate lane within the cir- culatory roadway, and the design of the exits also provides appropriate alignment to allow drivers to intuitively main- tain the appropriate lane. The report’s authors add that these alignment considerations often compete with the fastest path speed objectives. inscribed circle Diameter In the first FHWA Roundabouts Guide (Robinson et al. 2000), the authors state that the inscribed circle diameter (ICD) in a single-lane roundabout should be a minimum of 30 m (100 ft) to accommodate a WB-15 (WB-50) design vehicle. “Smaller roundabouts can be used for some local street or collector street intersections, where the design vehi- cle may be a bus or single-unit truck. At double-lane round- abouts, accommodating the design vehicle is usually not a constraint. The size of the roundabout is usually determined either by the need to achieve deflection or by the need to fit the entries and exits around the circumference with reason- able entry and exit radii between them.” Thus, the authors recommended that the ICD of a double-lane roundabout gen- erally be a minimum of 45 m (150 ft). The second edition of the FHWA Guide modified the ICD recommendations from the first edition; the second edition’s typical ICD ranges are shown in Table 18. entry width According to the first FHWA Roundabouts Guide (Robinson et al. 2000), determining the entry width and circulatory roadway width involves a trade-off between capacity and safety. The design should provide the minimum width necessary for capacity and accommodation of the design vehicle in order to maintain the highest level of safety. Typical entry widths for single-lane FIGURE 10 Radial alignment of roundabout entries (Robinson et al. 2000).

42 entrances range from 4.3 to 4.9 m (14 to 16 ft); however, values higher or lower than this range may be required for site-specific design vehicle and speed requirements for critical vehicle paths. Where wider entries are required, this can be done in two ways: by adding a full lane upstream of the entrance and maintain- ing parallel lanes through the entry, or by gradually widening the approach through flaring. When used, the Guide states that flare lengths should generally be a minimum of 25 m (80 ft) in urban areas and 40 m (130 ft) in rural areas. The second edition of the Guide (Rodegerdts et al. 2010) revised the previous guidelines to state that typical entry widths for single-lane entrances range from 14 to 18 ft, which are often flared from upstream approach widths. However, values higher or lower than this range may be appropriate for site- specific design vehicle and speed requirements for critical vehicle paths. A 15-ft entry width is a common starting value for a single-lane roundabout. NCHRP Report 672 also states that care be taken with entry widths greater than 18 ft or for those that exceed the width of the circulatory roadway, as drivers may mistakenly interpret the wide entry to be two lanes when there is only one receiving circulatory lane. intersection sight Distance Concerning ISD at roundabout approaches, the first edition of the FHWA Guide (Robinson et al. 2000) recommended the use of a critical headway of 6.5 s to determine the appro- priate length of the conflicting leg of the sight triangle. It further recommended that designers “provide no more than the minimum required intersection sight distance on each approach. Excessive intersection sight distance can lead to higher vehicle speeds that reduce the safety of the intersec- tion for all road users (e.g., vehicles, bicycles, pedestrians).” The authors also stated that landscaping can be effective in restricting sight distance to the minimum requirements. NCHRP Report 672 (Rodegerdts et al. 2010) advised the use of a critical headway of 5.0 s, based on the critical headway required for passenger cars. The authors added that this value represented an interim methodology pending further research. superelevation Guidelines in the first FHWA Roundabouts Guide state that, in general, “a cross-slope of 2% away from the central island should be used for the circulatory roadway. This technique of sloping outward [was] recommended for four main reasons: • It promotes safety by raising the elevation of the central island and improving its visibility. • It promotes lower circulating speeds. • It minimizes breaks in the cross slopes of the entrance and exit lanes. • It helps drain surface water to the outside of the roundabout.” The outward cross-slope design means vehicles making through and left-turn movements must negotiate the round- about at negative superelevation; however, the slow speeds through the circulatory roadway were generally expected to negate the effects of the slope on drivers. safety Researchers on NCHRP 3-65 (Rodegerdts et al. 2007) found that crash experience at selected intersections in the United States that had been converted to roundabouts showed an overall reduction in crash frequency; there were selected intersections at which this was not the case (e.g., either no change or a small increase in crash frequency), but in most cases, the crash counts at those locations were too small for increases to be statistically significant. In com- paring crash frequency to geometry, researchers listed these findings: • Eight of the ten sites with the lowest crash frequencies were single-lane roundabouts. • Twenty-six of the 30 sites with the lowest crash fre- quencies were single-lane roundabouts. • Two of the ten sites with the highest crash frequencies were single-lane roundabouts. • Nine of the 30 sites with the highest crash frequencies were single-lane roundabouts. Roundabout Configuration Typical Design Vehicle Common Inscribed Circle Diameter Range* Mini-Roundabout SU-30 45–90 ft Single-Lane Roundabout B-40 90–150 ft WB-50 105–150 ft WB-67 130–180 ft Multilane Roundabout (2 lanes) WB-50 150–220 ft WB-67 165–220 ft Multilane Roundabout (3 lanes) WB-50 200–250 ft WB-67 220–300 ft Source: Rodegerdts et al. (2010). *Assumes 90-degree angles between entries and no more than four legs. List of possible design vehicles is not all-inclusive. TABLE 18 TYPICAL INSCRIBED CIRCLE DIAMETER RANGES

43 • Crash frequency increased as the inscribed circle diam- eter increased, and as the number of vehicles entering the roundabout increased. • Crash frequency increased slightly as the number of legs to the roundabout increased. A review of multilane roundabout characteristics led the research team to believe that most sites were not designed using the natural vehicle path concept, which was likely because the design was completed before the introduction of the concept in the first edition of the FHWA Roundabout Guide. Lane widths also appeared to have an effect on crash frequency, particularly lanes that were narrower than those recommended by FHWA. Analysis of roundabouts in the United States led research- ers to conclude that the general principle was true that as the width of an entry increases, the capacity of the entry increases, while the safety of the entry decreases. However, extending the principle beyond the number of lanes to the actual entry width did not appear to have as strong a relation- ship in the United States as in other countries. Researchers suggested that, although the overall relationship between capacity and entry width appeared to hold true in terms of the aggregate number of lanes on the approach, changes in entry width within a single-lane entry has a much lower-order effect on capacity. The NCHRP 3-65 research team also reported that the angle between legs of a roundabout appeared to have a direct influence on entering-circulating crashes. As the angle to the next leg decreased, the number of entering-circulating crashes increased, suggesting that roundabouts with more than four legs or with skewed approaches tended to have more entering- circulating crashes. Analysis of U.S. data did not find a significant relationship between the capacity of the entry and the width of the splitter island, nor with the percent- age of exiting vehicles. The research team believed that as drivers became more comfortable and efficient in driving roundabouts the effect of the width of the splitter island and/or percentage of exiting vehicles could become more noticeable and recommended further study of the subject in the future. Researchers suggested that the critical headway estimate of 6.5 s in the first edition of the FHWA Roundabout Guide appeared to be somewhat conservative for design purposes for both single-lane and multilane entries. They recom- mended a lower value of 6.2 s for design purposes, which was approximately one standard deviation above the mean observed critical headway (Rodegerdts et al. 2007). Isebrands (2009) conducted a review of crashes at 17 inter- sections on rural high-speed roadways that were converted to roundabouts between 1993 and 2006; “high-speed” was defined as having a posted speed limit of 40 mph or greater. The number of years of before data averaged 4.6, with a minimum of 2.5 years and a maximum of 6.6 years. The after data showed greater variation, with an average of 5.5 years, a minimum of 1.8 years, and a maximum of 12.7 years of data. The results of her analysis showed 52% and 67% reductions in total crashes and crash rate, respectively. Moreover, the findings showed an 84% reduction in injury crashes and an 89% reduction in the injury crash rate. No fatal crashes occurred since the roundabouts were con- structed, compared with 11 fatal crashes that were reported in the before period. The number of angle crashes was also reduced by 86%. Pedestrian considerations The first FHWA Roundabout Guide (Robinson et al. 2000) discussed considerations for nonmotorized users. The authors provided design dimensions from Pein (1996) for key round- about design features, which are largely repeated in the second Roundabout Guide and are included in Table 19. The Guide added that “pedestrian crossing locations must balance pedes- trian convenience, pedestrian safety, and roundabout opera- tions.” With those issues in mind, the Guide recommended TABLE 19 KEY DIMENSIONS OF NONMOTORIzED ROUNDABOUT DESIGN USERS User Dimension (ft) Affected Roundabout Features Bicycle Length 5.9 Splitter island width at crosswalk Minimum operating width 4.0 Bicycle lane width Lateral clearance on each side 2.0 Shared bicycle–pedestrian path width 3.3 to obstructions Pedestrian (walking) Width 1.6 Sidewalk width, crosswalk width Wheelchair Minimum width 2.5 Sidewalk width, crosswalk width Operating width 3.0 Sidewalk width, crosswalk width Person pushing stroller Length 5.6 Splitter island width at crosswalk Skater Typical operating width 6.0 Sidewalk width Source: Pein (1996).

44 that pedestrian crossings be designed with the following characteristics: • “The pedestrian refuge should be a minimum width of 1.8 m (6 ft) to adequately provide shelter for persons pushing a stroller or walking a bicycle. • At single-lane roundabouts, the pedestrian crossing should be located one vehicle-length (7.5 m [25 ft]) away from the yield line. At double-lane roundabouts, the pedestrian crossing should be located one, two, or three car lengths (approximately 7.5 m, 15 m, or 22.5 m [25 ft, 50 ft, or 75 ft]) away from the yield line. • The pedestrian refuge should be designed at street level, rather than elevated to the height of the splitter island. This eliminates the need for ramps within the refuge area, which can be cumbersome for wheelchairs. • Ramps should be provided on each end of the crosswalk to connect the crosswalk to other crosswalks around the roundabout and to the sidewalk network. • It is recommended that a detectable warning surface, as recommended in the Americans with Disabilities Act Accessibility Guidelines (ADAAG), be applied to the surface of the refuge within the splitter island.” The second edition of the Guide recommended minimum splitter island dimensions, as shown in Figure 11, and the authors encouraged use of standard AASHTO island design for key dimensions, such as offset and nose radii. For side- walks, authors advised a setback distance of 1.5 m (5 ft), with a minimum of 0.6 m (2 ft). Research on NCHRP Project 3-65 included the review of pedestrian crossing activity at 10 legs on seven roundabout study sites (Rodegerdts et al. 2007). The researchers deter- mined that the majority of the 769 observed crossing events involved no interaction with a motor vehicle, where inter- action is defined as the pedestrian either accepting or reject- ing a gap when a vehicle was present. For those pedestrians who did interact with vehicles and ultimately crossed the leg, researchers categorized their behaviors as Normal, Hesitates, or Runs. Three categories of motorist yielding behavior were identified as well: • Active yield: The motorist slowed or stopped for a cross- ing pedestrian or a pedestrian waiting on the curb or split- ter island to cross. The pedestrian was the only reason the motorist stopped or slowed. • Passive yield: The motorist yielded to the pedestrian but was already stopped for another reason. This situation occurred most often when there was a queue of vehicles waiting to enter the roundabout or when the vehicle was already stopped for a prior pedestrian crossing event. • Did not yield: The motorist did not yield to a crossing pedestrian or a pedestrian waiting on the curb or splitter island to cross. Researchers determined that for pedestrians initiating a crossing on the entry side of one-lane sites, 15% of motorists did not yield to the pedestrian on either the entry or exit side. The remainder of the exit-side vehicles actively yielded. The remainder of the entry-side vehicles included 20% that were classified as passively yielding. For two-lane sites, the per- centage of nonyielding vehicles increased to 33% on the entry side and 45% on the exit side. For those vehicles that did yield, 9% and 2% were classified as passive yield for the entry and exit sides, respectively. When crossing began on the exit side, yielding improved for entry-side drivers but declined for exit-side drivers. In all categories, yield- ing at two-lane sites was lower than at one-lane sites. On average across all sites, approximately 30% of the motorists FIGURE 11 Minimum splitter island dimensions (Rodegerdts et al. 2010).

45 did not yield to pedestrians who were crossing or waiting to cross, although in all but one case the pedestrians were waiting to cross, so there was no imminent risk identified by the research team. Researchers also observed only four con- flicts in the 769 crossing events. Comparison with findings from a separate FHWA study (Carter et al. 2005b) indicated that driver yielding at roundabouts was better than at uncon- trolled approaches, but not as high as at stop signs or traffic signals. The researchers suggested that design changes could include reductions in exit radii, reductions in lane widths, and/or relocation of crosswalks (Rodegerdts et al. 2007). bicycle considerations The first FHWA Roundabout Guide (Robinson et al. 2000) recommended that the “designer should strive to provide bicyclists the choice of proceeding through the roundabout as either a vehicle or a pedestrian.” The Guide stated that, “in general, bicyclists are better served by treating them as vehicles; however, the best design provides both options to allow cyclists of varying degrees of skill to choose their more comfortable method of navigating the roundabout.” According to the Guide, to “accommodate bicyclists trav- eling as vehicles, bike lanes should be terminated in advance of the roundabout to encourage cyclists to mix with vehicle traffic.” Under this treatment, it was recommended that bike lanes end 100 ft upstream of the yield line to allow for merg- ing with vehicles. This method is most successful at smaller roundabouts with speeds below 20 mph, where bicycle speeds can more closely match vehicle speeds. To accommodate bicyclists who prefer not to use the cir- culatory roadway, the Guide advised that “a widened side- walk or a shared bicycle/pedestrian path may be provided physically separated from the circulatory roadway [i.e.,] not as a bike lane within the circulatory roadway. Ramps or other suitable connections [could] then be provided between this sidewalk or path and the bike lanes, shoulders, or road surface on the approaching and departing roadways.” The designer was advised to exercise care in locating and design- ing the bicycle ramps so that they are not misconstrued by pedestrians as unmarked pedestrian crossings, nor should the exits from the roadway onto a shared path allow cyclists to enter the shared path at excessive speeds. The second edition of the Roundabouts Guide (Rodegerdts et al. 2010) advises that, for nonmotorized users, one impor- tant consideration during the initial design stage is to main- tain or obtain adequate right-of-way outside the circulatory roadway for the sidewalks. All nonmotorized users who are likely to use the sidewalk regularly, including bi cyclists in situations where roundabouts are designed to provide bicycle access to sidewalks, should be considered in the design of the sidewalk width. Report authors recommended that bicycle lanes not be provided through the roundabout and be ter- minated upstream of the entrance line. They recommended designs that encourage bicycle users to merge into the gen- eral travel lanes and navigate the roundabout as a vehicle, explaining that the typical vehicle operating speed within the circulatory roadway is in the range of 15 to 25 mph, which is similar to that of a bicycle. Because multilane roundabouts are more challenging for bicyclists, additional design features may be appropriate for those locations. innOvative Designs A number of new, innovative, or otherwise unique intersection designs were topics of considerable attention between 2000 and 2010. An FHWA study by Hughes et al. (2010) examined four alternative intersection designs, reviewing characteristics related to geometric design, access management, traffic con- trol devices, and other features. The four designs included Dis- placed Left-Turn (DLT), Median U-Turn (MUT), Restricted Crossing U-Turn (RCUT), and Quadrant Roadway (QR) inter- sections. Findings from that study and others related to the geo- metric design of those intersection types are summarized here. Displaced left-turn The main feature of the DLT alternative intersection is the relocation of the left-turn movement on an approach to the other side of the opposing roadway, which consequently eliminates the left-turn phase for this approach at the main intersection (Hughes et al. 2010). Traffic that would nor- mally turn left at the main intersection first crosses the oppos- ing through lanes at a signal-controlled intersection several hundred feet upstream of the main intersection. Left-turning vehicles then travel on a new roadway parallel to the oppos- ing lanes and execute the left-turn maneuver simultaneously with the through traffic at the main intersection. The dashed line in Figure 12 illustrates a typical left-turn maneuver at a DLT intersection. The layout in Figure 12 is for a full ver- sion, which has DLT movements on all four approaches; after the eastbound vehicles turn northbound, they must travel through another crossover for southbound left-turning vehicles. This design reflects a shift of the through traffic lanes into the median in an attempt to minimize the need for additional right-of-way. At several locations where DLT intersections have been implemented as a retrofit to an exist- ing conventional at-grade intersection, the existing median has been preserved, and there is no shift in the through lanes. DLT can also be installed at a three-legged intersection with the displacement on the major road in only one direction. A study by Jagannathan and Bared (2005) investigated the design and operational performance of the DLT, then also known as the crossover displaced left-turn (xDL) or the continuous-flow intersection. The researchers’ purpose was to provide a simplified procedure to evaluate the DLT’s traffic performance and compare it with conventional intersections. Using microsimulation, they modeled typical geometries

46 over a wide distribution of traffic flow conditions for three different design configurations or cases. They concluded that their comparisons with conventional intersections showed considerable savings in average control delay and average queue length, as well as an increase in intersection capacity. They also concluded that their models provided an accessible tool for the practitioner to assess average delay and average queue length for those configurations. Simmonite and Chick (2005) conducted a similar study in the United Kingdom to evaluate a displaced right-turn intersection. They concluded that (1) intersection capacity can increase with a footprint similar to a large roundabout and only a small increase in costs, and (2) pedestrians and cyclists can easily be provided for without compromising the capacity using “Walk with Traffic” facilities. They suggested that the concept was an appropriate intersection type for use on the U.K. highway network, providing operational benefits where there were heavy right turns, full provision for non- motorized users, and an expected accident record unlikely to differ from other large signalized intersections. Hughes et al. (2010) stated that removal of conflict between the left-turn movement and the oncoming traffic at the main intersection is the primary design element in a DLT intersection. The DLT vehicles typically cross the opposing through traffic approximately 300 to 400 ft upstream of the main intersection under the control of another traffic signal. Research referenced in the report indicated that the appropri- ate upstream distance is dependent on queuing from the main intersection and on costs involved in constructing a left-turn storage area for the crossed-over left-turn movement. Radii of the crossover movements can range from 150 to 200 ft, whereas the radius of the next left-turn movement at the main intersection is dependent on the turning movement of the design vehicle. Lane widths at the crossover reverse curve need to be wider than 12 ft to accommodate larger design vehicles. Consideration could also be given to having wider lane widths (e.g., up to 15 ft) for the receiving crossroad. The angle between the DLT intersection left-turn lanes and the main through lanes is referred to as the crossover angle and is influenced by the median width and the alignment of the mainline lanes; a recommended range of values for this angle is 10 to 15 degrees. To minimize the footprint of the intersection, Hughes et al. (2010) stated that median widths can be reduced, but they still need to be adequate to accommodate signs. Designers are referred to the Green Book for minimum median widths, but caution is advised to also take into account the possi- bility of installing post-mounted signs in these medians for safe and effective channelization of traffic. Offsets for signs should be in accordance with the MUTCD. Results from an analysis by El Asawey and Sayed (2007) indicated that the capacity of a xDL intersection was higher than that of a conventional intersection by about 90%, and it outperformed conventional and upstream signalized cross- over intersections under all of their unbalanced-volume scenar- ios. They concluded that, for locations where right-of-way is not an issue, the xDL will be recommended for implementa- tion because of its superior performance compared with the other two intersections. median u-turn The MUT has been used in Michigan and other states as a treatment to balance intersection congestion and safety prob- lems (Hughes et al. 2010). The MUT intersection design involves the elimination of direct left turns from major and/ or minor approaches (usually both). Drivers desiring to turn left from the major road onto an intersecting cross street must first travel through the at-grade main intersection and then execute a U-turn at the median opening downstream of the intersection. These drivers then turn right at the cross street. Drivers on the minor street desiring to turn left onto the major road must first turn right at the main intersection, execute a U-turn at the downstream median opening, and proceed back through the main intersection. Figure 13 shows the left-turn movements of a typical MUT geometric design. The opti- mum directional crossover spacing was recommended to be 660 ft (±100 ft) from the main intersection. Elimination of left-turning traffic from the main intersection simplifies the signal operations at the intersection, which accounts for most of the intended benefits. The MUT intersection is typically a corridor treatment applied at signalized intersections. How- FIGURE 12 Left-turn movement on a typical DLT intersection approach (eastbound to northbound).

47 ever, the concept has also been used at isolated intersections to alleviate specific traffic operational and safety problems. The FHWA report states that the MUT intersection per- formed well on arterials that have sufficient median width to accommodate the U-turn maneuver. Because of Michi- gan experience with these intersections, the report discussed typical design values from the Michigan DOT. In general, Michigan corridors with MUT intersections have median widths ranging from 60 to 100 ft. This design is used as a cor- ridor treatment in Michigan, although it has also been used for isolated intersections. At an MUT, the design of the main intersection is similar to the design of a conventional intersection, except that the main intersection is designed for larger volumes of right-turn movements than a conventional intersection serving the same total volumes because the left-turning vehicles become right- turning vehicles. With this in mind, the intersection must be designed with right-turn bays of sufficient width and length to accommodate the volume of turning vehicles. Depending on the right-turn volume, dual right-turn lanes or an exclusive right-turn lane and an adjacent shared-use through and right- turn lane may be needed. Channelized right turns at an MUT intersection are rarely used, because they may require even more right-of-way, present a multistage pedestrian crossing, and create a more difficult driving maneuver for a driver turning right from the minor street and weaving over to use the U-turn crossover. At some MUT intersections (e.g., at partial MUT intersections), left turns from the side road are allowed as well as left-turn bays provided on the minor road approaches. The MUT intersection has secondary intersections at each of the crossover locations. One-way crossovers with deceleration/storage lanes are highly recommended. MDOT has developed design guidelines for directional median crossovers. In Michigan, the report states, it is cus- tomary for drivers of passenger vehicles to queue side-by- side in a 30-ft wide crossover and treat it as if it had two lanes. However, large trucks and other heavy vehicles typi- cally use the entire width of the crossover. MDOT uses striped two-lane crossovers (with two lanes of storage lead- ing up to the crossover) in some places. These crossovers are typically 36 ft wide. The FHWA report refers to the Green Book for minimum median widths, and it presents alterna- tives for locations with restricted right-of-way. restricted crossing u-turn Hughes et al. (2010) refer to RCUT intersections as a prom- ising solution for arterials with more dominant flows on the major road. Also referred to as superstreet intersections, they are described as having the potential to move more vehicles efficiently and safely than roadways with comparable traffic volumes that have conventional at-grade intersections with minimal disruptions to adjacent development. The RCUT intersection redirects left-turn and through movements from the side street approaches. Instead of allowing those move- ments to be made directly through the intersection, as in a conventional design, an RCUT intersection accommodates those movements by requiring drivers to turn right onto the main road and then make a U-turn maneuver at a one-way median opening 400 to 1,000 ft downstream. Figure 14 shows a conceptual diagram of an RCUT intersection. This configu- ration shown is generally intended for higher-volume major roads in suburban and rural areas, especially at intersections with relatively low through traffic volumes entering from the side road. For this type of intersection, left turns from the main road are similar to conventional intersections, made from left-turn lanes on the main road directly onto the side road. For this type of RCUT intersection design, pedestrians cross the main street in a diagonal fashion, going from one corner to the opposite corner. An RCUT design that does not permit direct left-turns is shown in Figure 15; this design channels all turning traffic to the crossovers on either side of the intersection. The key difference between an MUT intersection and an RCUT intersection is that an MUT intersection allows through movements from the side street. An RCUT intersection has either no median openings at the intersection or has one-way directional median openings to accommodate traffic making left turns from the main street onto the side street. Similar to the MUT intersection, the median width is a crucial design element for an RCUT intersection. The report states that desirable right-of-way widths needed to accommodate large FIGURE 13 MUT left-turn movements (based on Hughes et al. 2010). a) Major Street Movements Major street M in or s tre et b) Minor Street Movements Major street M in or s tre et

48 trucks without allowing vehicles to encroach on curbs or shoulders, assuming 12-ft-wide lanes and 10 ft of shoulder, range from approximately 140 ft for four-lane arterials to approximately 165 ft for eight-lane arterials. For this same situation, desirable minimum median widths between 47 and 71 ft are typically needed. As with MUT intersections, design- ers are referred to the Green Book for specific design guide- lines for minimum median widths, and much of the guidance in the FHWA report for crossover spacing for MUTs also is applied to RCUT intersections. The report states that several factors should be considered when selecting the appropriate spacing from a main intersec- tion to a U-turn crossover. Longer spacing between the main intersection and crossovers decreases spillback probabili- ties, providing more time and space for drivers to maneuver into the proper lane and read and respond to highway signs. Shorter spacing between the main intersection and cross- overs translates into shorter driving distances and travel times. AASHTO recommends spacing from 400 to 600 ft for MUT designs based on signal timing, whereas MDOT established 660 ± 100 ft as the standard spacing (Hughes et al. 2010). Quadrant roadway According to FHWA, the primary objective of a QR inter- section is to reduce delay at a severely congested intersection of two busy suburban or urban roadways and to reduce over- all travel time by removing left-turn movements (Hughes et al. 2010). A QR intersection can reportedly provide other benefits as well, such as improving pedestrian crossing time, and a QR intersection can be among the least costly of the four alternative intersections to construct and maintain. At a QR intersection, all four left-turn movements at a conventional four-legged intersection are rerouted to use a connector roadway in one quadrant. Figure 16 shows the connector road and how all four of the left-turning move- ments are rerouted to use it. Left turns from all approaches are prohibited at the main intersection, which consequently allows a simple two-phase signal operation at the main inter- section. Each terminus of the connector road is typically sig- nalized. These two secondary signal-controlled intersections usually require three phases. Key features in the geometric design of a QR intersection are choosing a quadrant in which to locate the connecting roadway; determining the number of connecting roadways; and designing the main intersection, the secondary inter- sections, and the horizontal alignment and cross section of the connecting road. In choosing a quadrant for the connecting roadway, common considerations are available right-of-way, construction cost, and effect on left-turn movements; for the latter, the quadrant is typically placed so that the movement with the highest volume is least affected by the new design of the intersection. That is, the left turn with the highest demand is the one that receives the most direct path. Discussion of Main street S id e st re et FIGURE 14 Conceptual RCUT configuration with direct left turns from the major road (based on Hughes et al. 2010). Si de s tre et Main street FIGURE 15 Basic RCUT intersection with no direct left turns (based on Hughes et al. 2010).

49 QR designs with multiple quadrant connectors is also pro- vided in the FHWA report. For the main intersection, the design would be similar to that of a conventional intersection with turn prohibitions. Appropriate pavement markings or median designs are to be employed to convey the message to drivers that no left turns or U-turns are allowed. Right-turn lane criteria are the same for a QR intersection as a conventional intersection except for the right turns in the quadrant with the connect- ing roadway. Right-turn demands do not change at the main intersection in the other three quadrants. Through volumes at the main intersection are higher in all four directions than at a conventional intersection because of rerouted left-turning traffic. Pedestrian crosswalks would normally be provided across all four approaches at the main intersection. Hughes et al. (2010) state that the distance from the main intersection to the secondary intersections is critical to the success of a QR intersection design. The considerations and trade-offs are similar to those between the main intersection and U-turn crossovers for an MUT or RCUT intersection. The distance needs to be sufficient to provide adequate vehi- cle storage and prevent spillback from one signal-controlled intersection to the next. It is also necessary to provide enough spacing for adequate signing and to ensure that each set of signal controls is visible. Longer distances lead to higher costs for right-of-way, construction, and maintenance of the connecting road. Longer distances may restrict progres- sion from one signal to the next on the main streets and can translate into more vehicle-hours of travel. Considering all of those factors, a minimum spacing of 500 ft from the center of the main intersection to the center of the secondary inter- sections is presented as adequate for many situations. The horizontal alignment of the connecting roadway is key to providing proper access to both roadways as well as any driveway connections. The authors recommend using the relevant geometric design data from the AASHTO Green Book for a design speed of 30 mph to determine the appropri- ate superelevation, radius, and runoff length. synchronized split-Phasing/ Double crossover intersection Bared et al. (2005) studied the operational characteristics of a synchronized split-phasing intersection, also called a double crossover intersection (DxI). An example of a DxI is shown in Figure 17. In this example, eastbound traffic crosses over to the left side at signalized Intersection A (small circle on the left of the figure), whereas the right-turners use the dedicated right lane before reaching A. The crossed traffic will cross over back to the right side at signalized Intersection C (small circle on the right). Westbound traffic also crosses over in a similar way. At Intersection B (large circle in the center of the figure), there is one through lane and one shared (through and left-turn) lane. No dedicated left-turn lanes are provided. Right-turn lanes are required for eastbound and westbound traffic. Merging lanes for the northbound and southbound right-turn movements are required. Radii of crossover move- ments can range from 150 to 200 ft, and the radius of the left-turn movement at B is 100 ft. Movements can be better understood by following the arrow markings in the figure. The northbound and the southbound traffic are similar to the cor- responding movements at a conventional intersection, with one left-turn lane, one through lane, and one shared (through plus right-turn) lane. The length of left-turn lane is 450 ft. FIGURE 16 Left-turn movements at a QR intersection (based on Hughes et al. 2010). a) Left Turn Pattern From The Arterial Arterial Cr os s st re et Q u ad ra n t roadway b) Left Turn Pattern From A Cross Street Arterial Cr os s st re et Q u ad ra n t roadway FIGURE 17 Double crossover intersection (Bared et al. 2005).

50 Results of traffic simulation were presented that showed input flow and throughput for DxI were similar, whereas the throughput was approximately 1,000 veh/hr lower than input flow for a conventional design. For peak volumes, the aver- age delay per vehicle for conventional design was 219 s/veh, compared with 87 s/veh for the DxI. The authors also noted that the numbers of stops, average stop time per vehicle, average queue, and maximum queue length were lower for the DxI than the conventional design. Finally, they con- cluded that including a pedestrian phase for the DxI pro- duced lower delay than the conventional intersection, and that the left-turn capacity in a DxI was more than twice that of a conventional design. A variation on the DCI is the Upstream Signalized Cross- over (USC) intersection. Sayed et al. (2006) investigated signal optimization strategies for USC intersections and identi- fied selected operational issues. They used microsimulation to model and analyze a USC intersection and, for compari- son, a conventional intersection. Their analysis revealed that, for relatively balanced volumes, a USC intersection could significantly reduce average vehicle delays, particularly when the volumes entering the intersection are relatively high. Additionally, the capacity of the USC intersection was found to be approximately 50% greater than that of a conventional intersection with similar geometry under balanced traffic volumes. For highly unbalanced volumes, particularly when the intersection volumes were relatively low, they found that a conventional intersection outperformed the USC inter- section. Overall, they concluded, the USC intersection showed considerable potential for situations in which one or more of the following conditions existed: (1) intersection volumes were balanced and near or over the capacity of a conven- tional intersection; (2) traffic volumes were somewhat unbal- anced, but the overall entering volumes were too high to be accommodated with a conventional intersection; or (3) the intersection had heavy left-turn volumes that caused excessive delays. El Asawey and Sayed (2007) also concluded that the capacity of a simulated USC intersection was approximately 50% higher than that of a conventional intersection. They noticed that with an increase in left-turn percentage from 20% to 30%, there was a relatively constant increase in delay for the USC, between 1 and 4 seconds. arterial interchange Eyler (2005) discussed a family of interchange designs that were developed for arterial roadways: split-level single- point, left-hand windmill, hybrid (half single-point, half windmill), and partial cloverleaf. Each of the four designs had one consistent requirement, that each at-grade inter- section in the interchange was the junction of only one turn- ing movement and one through movement. There was never a location where traffic crossed both directions of an inter- secting roadway. Eyler evaluated a selection of design varia- tions using VISSIM modeling, and he determined that the overall capacity was near 75% of a four-lane freeway. He also conducted a generalized cost comparison and found that while the new designs were more expensive than traditional at-grade intersections, the annual travel time savings would offset the construction cost within three years. His conclu- sion was that these designs merited further consideration as alternatives to both conventional at-grade intersections and typical expansion of arterials to freeways. alternatives for turning movements at rural intersections The purpose of NCHRP Project 15-30 (Maze et al. 2010) was to investigate alternative safety improvements at rural expressway intersections, to identify their relative effective- ness (if data were available), and to report any experiential information from those agencies that have tried the alterna- tives. After reviewing existing guidance and literature, the research team conducted case studies to investigate and docu- ment the effectiveness of ten treatments. Using that informa- tion, researchers recommended improvements to rural median intersection design guidance provided in the Green Book and the MUTCD for high-speed (50 mph or greater) expressways (divided highways with partial or no access control). The research team recommended that the next update to the Green Book include design guidance for rural expressway intersec- tion designs that eliminate or reduce far-side conflict points (e.g., J-turn intersections and offset T-intersections) or those that address the issue of gap selection for minor road drivers (e.g., left-turn median acceleration lanes and offset right-turn lanes). Examples of those designs are shown in Figure 18. Studies of the J-turn and offset-T designs revealed reduc- tions in crashes between 40% and 92%. Definitive results from the turning movement accommodations were unavail- able because of a small set of crash data, but the authors were optimistic about the treatments’ ability to reduce preventable crashes. two-level signalized intersection Shin et al. (2008) presented an unconventional intersection design (used in China) known as the Two-Level Signalized Intersection (TLSI) that completely separated east-west and north-south traffic. The TLSI, shown in Figure 19, also enabled the use of directional separation and leading, lagging, or over- lapping lefts on both upper and lower levels. They described the TLSI as a design consisting of two independently operat- ing intersections that is able to operate signals with flexibility according to changing traffic conditions. The results from their simulation modeling indicated that, compared with other innovative intersection types, the TLSI had the shortest delay times in most evaluation scenarios as well as the least sensitivity to variations in traffic volume.

51 However, the TLSI showed significant delay when traffic volumes on the major and minor roads are significantly dif- ferent, and it operated most efficiently when the two crossing roads had similar volumes of traffic. PeDestrian anD bicycle facilities According to NCHRP Report 500, Volume 18 (Raborn et al. 2008), there are several ways to modify the geometry of an intersection to improve bicycle safety, including: • “Reducing the crossing distance for bicyclists. • Realigning intersection approaches to reduce or elimi- nate intersection skew. • Modifying the geometry to facilitate bicycle movement at interchange on-ramps and off-ramps. • Providing refuge islands and raised medians.” At path/roadway intersections, an overpass or underpass allows for uninterrupted flow for bicyclists and completely eliminates exposure to vehicular traffic. These grade-separated crossings can improve safety and are desirable at some loca- tions. However, because grade-separated crossings can be quite expensive, may be considered unattractive, can poten- tially become sites of crime or vandalism, and may even decrease safety if not appropriately located and designed, these types of facilities are primarily used as measures of last resort. The AASHTO Bicycle Guide (1999) provides guidance on the design of overpasses and underpasses. This strategy is related to Strategy 9.1 A5—Install Overpasses/Underpasses in NCHRP Report 500, Volume 10: A Guide for Reducing Colli- sions Involving Pedestrians (zegeer et al. 2004). FHWA’s Signalized Intersections: Informational Guide (Rodegerdts et al. 2004) advises that “pedestrian facilities should be provided at all intersections in urban and suburban areas. In general,” the authors say, “design of the pedestrian facilities of an intersection with the most challenged users in mind—pedestrians with mobility or visual impairments— should be done, and the resulting design will serve all pedes- trians well.” The Guide adds that the “ADA requires that new and altered facilities constructed by, on behalf of, or for the use of State and local government entities be designed and constructed to be readily accessible to and usable by indi- viduals with disabilities.” FHWA’s guidelines are based on the premise that “pedestrians are faced with a number of dis- incentives to walking, including centers and services located far apart, physical barriers and interruptions along pedestrian routes, a perception that routes are unsafe owing to motor vehicle conflicts and crime, and routes that are [aestheti- cally] unpleasing.” FHWA notes “key elements that affect a pedestrian facility that practitioners should incorporate into their design: • Keep corners free of obstructions to provide enough room for pedestrians waiting to cross. • Maintain adequate lines of sight between drivers and pedestrians on the intersection corner and in the cross- walk. • Ensure curb ramps, transit stops (where applicable), push- buttons, etc., are easily accessible and meet ADAAG design standards. • Clearly indicate the actions pedestrians are expected to take at crossing locations. • Design corner radii to ensure vehicles do not drive over the pedestrian area yet are able to maintain appropriate turning speeds. • Ensure crosswalks clearly indicate where crossings should occur and are in desirable locations. • Provide appropriate intervals for crossings and minimize wait time. • Limit exposure to conflicting traffic and provide refuges where necessary. • Ensure the crosswalk is a direct continuation of the pedestrian’s travel path. • Ensure the crossing is free of barriers, obstacles, and hazards.” FIGURE 18 Diagrams of median intersection designs on rural expressways (based on Maze et al. 2010). a) J-Turn Intersection b) Offset T-Intersection c) Left-Turn Median Acceleration Lanes d) Offset Right-Turn Lane

52 Where on-street bicycle lanes or off-street bicycle paths enter an intersection, FHWA (Rodegerdts et al. 2004) advises that intersection design should accommodate the needs of cyclists in safely navigating such a large and often compli- cated intersection. It is recommended that geometric features to be considered include: • Bike lanes and bike lane transitions between through lanes and right-turn lanes. • Left-turn bike lanes. • Median refuges with a width to accommodate a bicycle: 6 ft = poor; 8 ft = satisfactory; 10 ft = good. • Separate facilities if no safe routes can be provided through the intersection itself. “Curb ramps provide access between the sidewalk and roadway for people using wheelchairs, strollers, walkers, crutches, handcarts, bicycles, and also for pedestrians with mobility impairments who have trouble stepping up and down high curbs. Curb ramps must be installed at all inter- sections and midblock locations where pedestrian crossings exist, as mandated by federal legislation,” notably the 1990 Americans with Disabilities Act (ADA). According to the Proposed Accessibility Guidelines for Pedestrian Facilities in the Public Right of Way, curb ramps must have a slope of no more than 1:12 (must not exceed 1 in./ft), and a maximum slope on any side flares of 1:10 (United States Access Board 2011). Additional details of curb ramp design are also pro- vided in the Proposed Guidelines. Channelized turning lanes pose a potential risk to pedes- trians, particularly those with disabilities. Researchers on NCHRP Project 3-78 (Schroeder et al. 2011) found anec- dotal evidence that a crosswalk located in the middle of a turning lane is preferable to a crosswalk at the upstream or downstream portion of the turn lane. The middle crosswalk establishes a short crossing path roughly perpendicular to the trajectory of turning vehicles (useful for establishing pedes- trian alignment), and it physically separates the conflict of turning drivers and pedestrians with the downstream merge point. Based on turning radii and associated design speeds, they posited that this was the likely location where speeds of right-turning vehicles would be lowest. transit cOnsiDeratiOns TCRP sponsored a recent project to develop guidance for transit and highway agencies in the operations, planning, and functional designs of at-grade crossings of busways in FIGURE 19 Concept design of two-level signalized intersection (Shin et al. 2008). Upper Level Lower Level

53 physically separated rights-of-way by roadways, bike paths, or pedestrian facilities. TCRP Report 117 (Eccles et al. 2007) documents the activities on that project, and the guidance contained in that report is intended to “provide information that can be applied to enhance safety at busway crossings while maintaining efficient transit and highway operations and minimizing pedestrian delay.” General design principles and guidelines included: • Provide simple intersection designs. • Provide clear visual cues to make busway intersections conspicuous. • Maximize driver and pedestrian expectancy. • Separate conflicting movements. • Minimize street crossings. • Incorporate design features that improve safety for vul- nerable users. • Coordinate geometric design features and traffic control devices. TCRP Report 117 discussed four types of busways found at intersections: median busways, side-aligned busways, sep- arated right-of-way busways, and bus-only ramps. For each busway type, the report contains guidance on safety issues, basic geometry (including placement of bus stops), and traf- fic control, as well as an example of an intersection that uses each type of busway. Safety issues were generally related to the complexity and/or unfamiliarity of the arrangement of the intersection and the accommodation of pedestrians. Geometry guidelines pertained to channelization and control of turning movements to protect buses and passengers, provision of suf- ficient right-of-way to include the number of necessary travel lanes, and providing design consistency between busway lanes and the adjacent general-purpose lanes. The traffic con- trol device most frequently recommended was traffic signals; suggestions on timing and phasing were provided to promote optimization of capacity and safety. access management Recent research has established support for use of access management principles in improving intersection design and safety. The optimal situation is to avoid driveway conflicts before they develop (Neuman et al. 2003b). “This requires coordination with local land use planners and zoning boards in establishing safe development policies and procedures. Avoidance of high-volume driveways near congested or other- wise critical intersections is desirable. Driveway-permitting staff within highway agencies also needs to have an under- standing of the safety consequences of driveway requests.” Some recent research, findings, and discussion related to access management are contained in this section. zhou et al. (2002) studied the operational effectiveness of using right-turn-plus-U-turn (RTUT) as an alternative to DLTs from driveways where raised-curb medians were installed on six-lane highways at eight sites in Florida. After analyz- ing the data, the authors concluded that U-turns could have better operational performance than DLTs under certain traffic conditions, which they said implied that directional median opening designs would provide more efficient traffic flow than full median openings. They also stated that RTUT would provide better safety with regard to traffic conflicts and fewer effects on through-traffic operations of a major highway; they added that the majority of traffic on the major street in the study was in platoon flow because the signal spacing at study sites was less than 2 mi. Turns could not be made when the platoon was passing the driveway, and strag- glers and left-turn-in movements from major roads between the platoons affected the ability to make turns. Carter et al. (2005a) examined the operational and safety effects of U-turns at signalized intersections. The operational analysis involved measurements of vehicle headways in exclu- sive left-turn lanes at 14 signalized intersections. Regression analysis of saturation flow data showed a 1.8% saturation flow rate loss in the left-turn lane for every 10% increase in U-turn percentage and an additional 1.5% loss for every 10% U-turns if the U-turning movement was opposed by protected right- turn overlap from the cross street. The safety analysis involved a set of 78 intersections, 54 sites chosen randomly and 24 sites selected on the basis of their reputation as U-turn problem sites. Although the researchers used a group of study sites that was biased toward sites with high U-turn percentages, they found that 65 of the 78 sites did not have any collisions involv- ing U-turns in the 3-year study period. U-turn collisions at the remaining 13 sites ranged from 0.33 to 3.0 collisions per year. Sites with double left-turn lanes, protected right-turn overlap, or high left-turn and conflicting right-turn traffic volumes were found to have a significantly greater number of U-turn col- lisions. Researchers concluded that, overall, U-turns do not have the large negative effect at signalized intersections that many have assumed, as safety and operational effects were minimal. In NCHRP Project 17-21, researchers determined state and local agency design practices and policies related to unsig- nalized median openings for U-turns, such as those shown in Figure 20. After seven categories of midblock and inter- section median designs were identified, the research team assessed the designs’ effects on safety through field obser- vation and crash data analysis for 115 unsignalized median opening sites with both crash and field data. This knowledge was transferred into design guidelines and a methodology for comparing the expected safety performance of different designs, to enable engineers in setting policy, establishing project-level design, and discussing the impacts of medi- ans with business and property owners. As documented in NCHRP Report 524 (Potts et al. 2004), researchers made the following conclusions: • As medians are used more extensively on arterial high- ways, with direct left-turn access limited to selected

54 locations, many arterial highways experience fewer mid- block left-turn maneuvers and more U-turn maneuvers at unsignalized median openings. • Field studies at various median openings in urban arterial corridors found estimated U-turn volumes of no more than 3.2% of the major-road traffic volumes at those locations. At rural median openings, U-turn volumes were found to represent at most 1.4% of the major-road traffic volumes at those locations. • Accidents related to U-turn and left-turn maneuvers at unsignalized median openings occurred very infre- quently. The 103 median opening study sites on urban arterial corridors experienced an annual U-turn plus left-turn crash average of 0.41. Twelve median open- ings on rural arterial corridors had an annual average crash total of 0.20. Overall, at these median openings, U-turns represented 58% of the median opening move- ments and left turns represented 42%. Based on these limited crash frequencies, researchers concluded that there was no indication that U-turns at unsignalized median openings constituted a major safety concern. • For urban arterial corridors, median opening crash rates were substantially lower for midblock median openings than for median openings at three- and four-leg inter- sections. For example, the crash rate per million median opening movements (U-turn plus left-turn maneuvers) at a directional midblock median opening was typically only about 14% of the median opening crash rate for a directional median opening at a three-leg intersection. • Crash rates at directional median openings on urban arte- rial corridors were lower than at traditional median open- ings, and conventional three-leg median openings had lower crash rates than corresponding four-leg openings. • Where directional median openings were considered as alternatives to conventional median openings, two or more directional median openings were usually required to serve the same traffic movements as one conventional median opening. Therefore, researchers concluded that design decisions consider the relative safety and opera- tional efficiency of all directional median openings in comparison with the single conventional median opening. • Analysis of field data found that, for most types of median openings, most observed traffic conflicts involved major- road through vehicles having to brake for vehicles turning from the median opening onto the major road. • At urban unsignalized intersections, the research found that installation of a left-turn lane on one approach would be expected to reduce accidents by 27% for four-leg intersections and by 33% for three-leg intersections. • The minimum spacing between median openings then used by highway agencies ranged from 152 to 805 m (500 to 2,640 ft) in rural areas and 91 to 805 m (300 to 2,640 ft) in urban areas. In most cases, highway agen- cies used spacings between median openings in the upper end of these ranges, but there was no indication that safety problems resulted from occasional use of median opening spacings as short as 91 to 152 m (300 to 500 ft). Based on these and other findings, supplemented in part by conclusions in NCHRP Report 348 (Koepke and Levinson 1992), NCHRP Report 375 (Harwood et al. 1995), and NCHRP Report 420 (Gluck et al. 1999), the NCHRP 17-21 research team developed and presented a five-step methodology for comparing the expected safety performance of median open- ing design alternatives to assist in the selection of median opening types and the comparison of alternative median open- ing arrangements. As part of their conclusions, the research team recommended the following: • Unsignalized median openings may be used for a broad range of major- and minor-road traffic volumes. How- ever, if the major- and minor-road volumes exceed the traffic volumes given in the MUTCD signalization war- rants, signalization of the median opening needs to be considered. • The effects of U-turn and left-turn volumes on median opening crash frequency cannot be separated, because a review of crash data for median openings found that crash report data do not distinguish clearly between crashes involving U-turn maneuvers and those involv- ing left-turn maneuvers. • For rural unsignalized intersections: – They should have medians that are as wide as practical, as long as the median is not so wide that approaching vehicles on the crossroad cannot see both roadways of the divided highway. – Where the AASHTO passenger car is used as the design vehicle, a minimum median width of 8 m (25 ft) is recommended. – Where a large truck is used as the design vehicle, a median width of 21 to 31 m (70 to 100 ft) generally would be selected. If such a median width cannot be FIGURE 20 Median opening for left-turn lane on a four-lane divided suburban arterial at an unsignalized three-leg intersection (Credit: Dan Walker, Texas Transportation Institute).

55 provided, consideration should be given to provid- ing a loon. • For suburban unsignalized intersections: – Median widths at suburban unsignalized inter- sections generally should be as narrow as possible while providing sufficient space in the median for the appropriate left-turn treatment. – Median widths between 4.2 and 7.2 m (14 and 24 ft) will accommodate left-turn lanes, but are not wide enough to store a crossing or turning vehicle in the median. – Medians wider than 7.6 m (25 ft) may be used, but cross- road vehicles making turning and crossing maneuvers may stop on the median roadway. – Median widths of more than 15 m (50 ft) gener- ally should be avoided at suburban, unsignalized intersections. • Median opening lengths at rural divided highway inter- sections generally should be kept to the minimum pos- sible. Increases in median opening length were found to be correlated with higher rates of undesirable driving behavior. In contrast, researchers found no reason that the median opening in urban and suburban areas should not be as long as necessary. • Median opening spacing for rural areas typically ranged from 150 to 805 m (500 to 2,640 ft); a minimum median opening spacing of 150 m (500 ft) was recommended in rural areas. Typically, median opening spacing sub- stantially longer than 150 m (500 ft) was considered to be appropriate, unless two public road intersections or major driveways are located relatively close together. • Median opening spacing for urban areas typically ranged from 90 to 805 m (300 to 2,640 ft); a minimum median opening spacing of 90 m (300 ft) was recommended in urban areas. Researchers stated that, whenever practical, median opening spacing greater than 90 m (300 ft) should be used in urban areas. • U-turn maneuvers should not be encouraged at loca- tions with limited sight distance. Furthermore, sight distance is an important issue in determining locations where U-turns by larger vehicles should be permitted or encouraged. ISD based on the criteria in the AASHTO Green Book for Cases B1, B2, and F should be available to accommodate U-turns and left turns at unsignalized median openings. Gattis et al. (2010) presented guidelines for driveway spac- ing near intersections, both signalized and unsignalized. For unsignalized intersections, they stated that spacing should not interfere with safe and relatively unimpeded movement on the through roadway, and driveway spacing practices should provide reasonable access to abutting private property. Other general guidelines included: • The needed distance between successive connections (both driveways and side streets) increases with higher operating speeds, higher access classifications for the public roadway, and higher driveway volumes. • A driveway should not be located within the functional area of an intersection or in the influence area of the upstream and downstream driveways. • Left-turn lane storage requirements should be consid- ered when determining the driveway influence area and can limit how closely driveways can be spaced. • On roadways that are undivided or have TWLTLs, the alignment of driveways on opposite sides of the road needs to be considered. Driveways on opposite sides of a lower-volume roadway may be aligned across from each other. Alternatively, they should be spaced so that those drivers desiring to travel between the driveways on opposing sides of the roadway need to make a dis- tinct right turn followed by a left turn (or a left followed by a right). A much longer separation is needed on a higher-speed, higher-volume roadway. • On roadways with restrictive medians, the spacing between right-turn access points on opposite sides of the road can be treated separately. • Ideally, driveway access for a major development involv- ing left-turn egress movements should be located where effective coordination of traffic signals would be achiev- able if there is a need to signalize the driveway. • Driveway connections to public roadways are subject to the same intersection control device analyses as are street intersections. If existing or future volumes war- rant installing a traffic signal, and signalized spacing requirements cannot be met, left-turn access should be subject to closure in one or both directions. For driveways near signalized intersections, Gattis stated that the needed minimum separation distance (i.e., corner clearance) will depend on the function, operation, and design features of the roadway and the characteristics of the access connection, considering the basic principle of locat- ing one connection outside of the functional area of another connection. For a driveway upstream of or approaching a signalized location on a major road, the functional area was defined to include the PRT, maneuver distance, and storage length of the traffic on that approach. The recom- mended spacing would provide separation between the conflicting movements occurring at the signal and the conflicting movements occurring at the driveway. In addi- tion, this spacing would enable the driveway to operate without being obstructed by the traffic backing up from the signal. summary Of key finDings This section summarizes key findings from the research noted in this chapter. This is an annotated summary; conclu- sions and recommendations are those of the authors of the references cited.

56 intersection alignment • Avoid approach grades to an intersection of greater than 6%. On higher design speed facilities (50 mph and greater) a maximum grade of 3% should be considered (Rodegerdts et al. 2004). • Avoid locating intersections along a horizontal curve of the intersecting road (Rodegerdts et al. 2004). • Strive for an intersection platform (including sidewalks) with a cross slope not exceeding 2%, as needed for acces- sibility (Rodegerdts et al. 2004). • Approach curvature can be used as a treatment to force a reduction in vehicle speed through the introduction of horizontal deflection at high-speed intersection approaches, but it is discouraged at downhill approaches (Ray et al. 2008). • A skew angle greater than 20 degrees should not be used in design when the design vehicle is a large vehicle or semitrailer (Son et al. 2002). • A minimum skew angle of 15 degrees should be used to accommodate age-related performance deficits at intersections where right-of-way is restricted (Staplin et al. 2002). auxiliary lanes • Adding left-turn lanes is effective in improving safety at signalized and unsignalized intersections, reducing crashes between 10% and 44% (Harwood et al. 2002). • Positive results can also be expected for right-turn lanes, with reductions in total intersection crashes between 4% and 14% (Harwood et al. 2002). • A method was developed to identify where installation of right-turn lanes at unsignalized intersections and major driveways would be cost-effective, indicating combinations of through-traffic volumes and right-turn volumes for which provision of a right-turn lane would be recommended. The economic analysis procedure can be applied by highway agencies using site-specific values for ADTs, turning volumes, crash frequency, and con- struction cost for any specific location (or group of similar locations) of interest (Potts et al. 2007a). modern roundabouts • A series of projects during the decade (2000–2010) led to the publication of two FHWA Informational Guides con- taining recommendations and guidelines for all aspects of roundabout design. • General overarching principles of geometric design of roundabouts (Rodegerdts et al. 2010) included: – “Provide slow entry speeds and consistent speeds through the roundabout by using deflection. – Provide the appropriate number of lanes and lane assignment to achieve adequate capacity, lane volume balance, and lane continuity. – Provide smooth channelization that is intuitive to drivers and results in vehicles naturally using the intended lanes. – Provide adequate accommodation for the design vehicles. – Design to meet the needs of pedestrians and cyclists. – Provide appropriate sight distance and visibility for driver recognition of the intersection and con- flicting users.” • Maximum entering design speeds are based on a the- oretical fastest path of 20 to 25 mph for single-lane roundabouts and 25 to 30 mph for multilane round- abouts (Rodegerdts et al. 2010). • Roundabout alignment is described as “optimally located when the centerlines of all approach legs pass through the center of the inscribed circle” (Robinson et al. 2000). • Common inscribed circle diameters for single-lane roundabouts vary from 90 to 180 ft, depending on design vehicle (Rodegerdts et al. 2010). • Designers should provide no more than the minimum required ISD on each approach, [because] excessive ISD can lead to higher vehicle speeds that reduce the safety of the intersection for all road users (Robinson et al. 2000). • Crash experience at selected intersections in the United States indicates an overall reduction in crash frequency at intersections converted to roundabouts (Rodegerdts et al. 2007). • Pedestrian refuge should be a minimum width of 6 ft to adequately provide shelter for persons pushing a stroller or walking a bicycle (Robinson et al. 2000). • “At single-lane roundabouts, the pedestrian crossing should be located one vehicle-length (25 ft) away from the yield line. At double-lane roundabouts, the pedes- trian crossing should be located one, two, or three car lengths (approximately 25 ft, 50 ft, or 75 ft) away from the yield line” (Robinson et al. 2000). • “The pedestrian refuge should be designed at street level, rather than elevated to the height of the split- ter island. This eliminates the need for ramps within the refuge area, which can be cumbersome for wheel- chairs” (Robinson et al. 2000). • Ramps should be provided on each end of crosswalks to connect the crosswalk to other crosswalks around the roundabout and to the sidewalk network (Robinson et al. 2000). • A detectable warning surface, as recommended in the ADAAG, should be applied to the surface of the refuge within the splitter island (Robinson et al. 2000). • Use of standard AASHTO island design for key dimen- sions, such as offset and nose radii, is encouraged. For sidewalks, a setback distance of 5 ft, with a minimum of 2 ft is advised (Robinson et al. 2000). • For nonmotorized users such as bicyclists, one important consideration during the initial design stage is to main- tain or obtain adequate right-of-way outside the circula- tory roadway for the sidewalks. All nonmotorized users who are likely to use the sidewalk regularly, including

57 bicyclists in situations where roundabouts are designed to provide bicycle access to sidewalks, should be con- sidered in the design of the sidewalk width. Recom- mended designs for single-lane roundabouts encourage bicycle users to merge into the general travel lanes and navigate the roundabout as a vehicle, explaining that the typical vehicle operating speed within the circula- tory roadway is in the range of 15 to 25 mph, which is similar to that of a bicycle (Rodegerdts et al. 2010). innovative intersection Designs A number of new or innovative intersection designs were considered during the decade; each of the following was described in one or more studies. • Displaced Left Turns showed considerable savings in average control delay and average queue length, as well as an increase in intersection capacity, in one series of microsimulation analyses (Hughes et al. 2010). • Median U-turns are typically a corridor treatment applied at signalized intersections, but are also used at isolated intersections to alleviate specific traffic operational and safety problems (Hughes et al. 2010). • Median width of Restricted Crossing U-Turns is a cru- cial design element to accommodate large trucks with- out allowing vehicles to encroach on curbs or shoulders (Hughes et al. 2010). • Quadrant Roadways should be designed so that the left turn with the highest demand is the one that receives the most direct path (Hughes et al. 2010). • Double Crossover Intersections are found to have greater throughput than a conventional intersection, along with lower values for number of stops, average stop time per vehicle, average queue, and maximum queue length (Bared et al. 2005). • Arterial Interchanges have an overall capacity near 75% of a four-lane freeway (Eyler 2005). • J-Turn and Offset-T designs had reductions in crashes between 40% and 92% (Maze et al. 2010). • Two-Level Signalized Intersections produced mod- eled results with the shortest delay times in most evaluation scenarios as well as the least sensitivity to variations in traffic volume compared with other inno- vative intersection types; however, delay increased when flow was unbalanced between the two crossing roads (Shin et al. 2008). • The additional right-of-way needed to construct each of these innovative designs was mentioned as a potential drawback by every report and author that addressed the issue of the intersection’s footprint. Pedestrian and bicycle facilities at intersections • Suggested strategies (Raborn et al. 2008) for modifying intersections to accommodate bicycles and pedestrians included: – Reducing the crossing distance for bicyclists, – Realigning intersection approaches to reduce or eliminate intersection skew, – Modifying the geometry to facilitate bicycle move- ment at interchange on-ramps and off-ramps, – Providing refuge islands and raised medians, and – Grade-separated crossings. • “Pedestrian facilities should be provided at all inter- sections in urban and suburban areas. In general, design of pedestrian facilities with the most challenged users in mind—pedestrians with mobility or visual impairments—should be done, and the resulting design will serve all pedestrians well. ADA requires that new and altered facilities constructed by, on behalf of, or for the use of State and local government entities be designed and constructed to be readily accessible to and usable by individuals with disabilities” (Rodegerdts et al. 2004). • Practitioners should incorporate key elements that affect a pedestrian facility into their design (Rodegerdts et al. 2004): – “Keep corners free of obstructions to provide enough room for pedestrians waiting to cross. – Maintain adequate lines of sight between drivers and pedestrians on the intersection corner and in the crosswalk. – Ensure curb ramps, transit stops (where applicable), pushbuttons, etc., are easily accessible and meet ADAAG design standards. – Clearly indicate the actions pedestrians are expected to take at crossing locations. – Design corner radii to ensure vehicles do not drive over the pedestrian area yet are able to maintain appropriate turning speeds. – Ensure crosswalks clearly indicate where crossings should occur and are in desirable locations. – Provide appropriate intervals for crossings and mini- mize wait time. – Limit exposure to conflicting traffic, and provide ref- uges where necessary. – Ensure the crosswalk is a direct continuation of the pedestrian’s travel path. – Ensure the crossing is free of barriers, obstacles, and hazards.” transit considerations • General intersection design principles and guidelines for transit issues (Eccles et al. 2007) include: – “Provide simple intersection designs. – Provide clear visual cues to make busway inter- sections conspicuous. – Maximize driver and pedestrian expectancy. – Separate conflicting movements. – Minimize street crossings. – Incorporate design features that improve safety for vulnerable users.

58 – Coordinate geometric design features and traffic control devices.” • There are four types of busways found at intersections: median busways, side-aligned busways, separated right- of-way busways, and bus-only ramps. Each busway type has unique characteristics that are considerations for guidance on safety issues, basic geometry (includ- ing placement of bus stops), and traffic control, along with examples of appropriate intersections for each type of busway (Eccles et al. 2007). access management at intersections • Right-turn-plus-U-turn could have better operational performance than direct left turns under certain traffic conditions, implying that directional median opening designs could provide more efficient traffic flow than full median openings (zhou et al. 2002). • U-turns at signalized intersections resulted in a 1.8% saturation flow rate loss in the left-turn lane for every 10% increase in U-turn percentage and an additional 1.5% loss for every 10% U-turns if the U-turning move- ment was opposed by protected right-turn overlap from the cross street (Carter et al. 2005a). • Recommended practices (Potts et al. 2004) for rural unsignalized intersections include: – They should have medians that are as wide as practical, as long as the median is not so wide that approaching vehicles on the crossroad cannot see both roadways of the divided highway. – Where the AASHTO passenger car is used as the design vehicle, a minimum median width of 25 ft is recommended. – Where a large truck is used as the design vehicle, a median width of 70 to 100 ft generally should be selected. If such a median width cannot be provided, consideration should be given to providing a loon. • Recommended practices (Potts et al. 2004) for suburban unsignalized intersections include: – Median widths at suburban unsignalized inter- sections generally should be as narrow as possible while providing sufficient space in the median for the appropriate left-turn treatment. – Median widths between 14 and 24 ft will accommo- date left-turn lanes, but are not wide enough to store a crossing or turning vehicle in the median. – Medians wider than 25 ft may be used, but crossroad vehicles making turning and crossing maneuvers may stop on the median roadway. – Median widths of more than 50 ft generally should be avoided at suburban, unsignalized intersections. • Median opening lengths at rural divided highway inter- sections generally should be kept to the minimum pos- sible. Increases in median opening length are correlated with higher rates of undesirable driving behavior. In contrast, the median opening in urban and suburban areas can be as long as necessary (Potts et al. 2004).

Next: Chapter Six - Interchanges »
Recent Roadway Geometric Design Research for Improved Safety and Operations Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Synthesis 432: Recent Roadway Geometric Design Research for Improved Safety and Operations reviews and summarizes roadway geometric design literature completed and published from 2001 through early 2011, particularly research that identified impacts on safety and operations.

The report is structured to correspond to chapters in the American Association of State Highway and Transportation Officials’ A Policy on Geometric Design of Highways and Streets, more commonly referred to as the Green Book.

NCHRP Synthesis 432 is an update of NCHRP Synthesis 299 on the same topic published in 2001.

  1. ×

    Welcome to OpenBook!

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

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

    No Thanks Take a Tour »
  2. ×

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

    « Back Next »
  3. ×

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

    « Back Next »
  4. ×

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

    « Back Next »
  5. ×

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

    « Back Next »
  6. ×

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

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

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

    « Back Next »
Stay Connected!