Human Factors Guidelines for Road Systems: Second Edition (2012)

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.

Task Analysis of Curve Driving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-2 The Influence of Perceptual Factors on Curve Driving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-4 Speed Selection on Horizontal Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-6 Countermeasures for Improving Steering and Vehicle Control Through Curves . . . . . . . . . .6-8 Countermeasures to Improve Pavement Delineation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-10 Signs on Horizontal Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-12 6-1 C H A P T E R 6 Curves (Horizontal Alignment)

HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 6-2 TASK ANALYSIS OF CURVE DRIVING Introduction This guideline identifies the basic activities that drivers would ty pically perform while trying to safely navigate a single ho rizontal curve. This information is useful because (1) it can help identif y segments of the curve driving task that are more demanding a nd require the driver to pay closer attention to basic vehicle c ontrol and visual information acquisition, and (2) it identifies t he key information and vehicle control requirements in different parts of the curve driving task. This information has design implica tions because workload is influenced by design aspects such as design c onsistency , degree of curvature, and lane width. In particula r, identify ing high workload components of the curve driving task provides an indication of where drivers could benefit from havin g their driving tasks made easier to perform (e.g., clearer roadwa y delineation, wider lanes, longer radius), or benefit from the elimination of potential visual distractions. The figure and table below show the different curve segm ents, as well as key driving tasks and constraints. De si gn Gu id e lin es Be ca us e dr iv er s ha ve hi gher vi su al dema nd s du ri ng cu rv e entr y an d na vi ga ti on —e sp ec ia lly wi th sh ar p cu rv es —c ur ve s sh ou ld be de si gned to mi ni mi ze a ddi ti on al wo rk lo ad im po se d on dr iv er s. Dr iv er vi su al dema nd s ar e gr ea te st ju st be fo re an d du ri ng cu rv e entr y an d na vi ga ti on be ca us e dr iv er s ty pi ca lly sp end mo st of thei r ti me l ook in g at the i mmedi at e ro ad wa y fo r ve hi cl e gu id an ce in fo rm at io n. So me Ge ne ra l Im p lic at io ns fo r th e De si gn of Ho ri zo nt al Cu rv es • Av oi d pr es enti ng vi su a lly co mp le x in fo rm at io n (e .g ., th at re qu ir es re ad in g an d/ or in te rp re ta ti on ) wi th in 75 to 100 m or 4 to 5 s of the po in t of cu rv at ur e, or wi th in it . • Ke y na vi ga ti on an d gu id an ce in fo rm at io n, su ch as la ne ma rk in gs an d de lin ea to rs /r ef le ct or s, sh ou ld be cl ea rl y vi si bl e in pe ri ph er al vi si on , es pe ci a lly un de r ni ghtti me co nd it io ns . • Mi ni mi ze the pr es en ce of ne ar by vi su al st im u li th at ar e po tent ia lly di st ra ct in g (e .g ., si gn ag e/ ad ve rt is ements th at “p op out” or i rre gu la r/ unus ua l ro ad si de sc ener y/ fo lia ge ). • Vi su al dema nd s a ppe ar to be lin ea rl y re la te d to cu rv e ra di us and un re la te d to de fl ec ti on an gl e. Cu rv es wi th a cu rv at ur e of 9 degr ees or gr ea te r ar e hi gh ly dema nd in g re la ti ve to mo re gr ad ua l cu rv es . Based Primarily on Expert Judgment Based Equally on Expert Judgment and Empirical Data Based Primarily on Empirical Data 2. Cu rv e Di sc ov er y 3. En tr y an d Ne go ti at ion 4 . Ex it 1. A pp roa ch Tangent Point Po in t of Cu rv at ur e 75 - 100 m (˜ 4 se c) Ex pe ct an cy Ef fe ct s 1. A ppr oa ch 2. Cu rv e Di sc ov er y 3. En tr y an d Ne go ti at io n 4. Ex it Ke y Dr iv in g Ta sk s 1. 1 Lo ca te be nd 1. 2 Ge t av a ila bl e sp eed in - fo rm at io n fr om si gn ag e 1. 3 Ma ke in it ia l sp eed ad ju st me nt s 2. 1 De te rm in e cu rv at ur e 2. 2 A sse ss ro ad wa y co nd it io ns 2. 3 Ma ke a ddi ti on al sp eed ad ju st me nt s 2. 4 Ad ju st pa th fo r cu rv e entr y 3. 1 Ad ju st sp eed ba se d on cu rv at ur e/ la te ra l a cce le ra ti on 3. 2 Ma in ta in pr op er tr aj ec to ry 3. 3 Ma in ta in sa fe la ne po si ti on 4. 1 A cce le ra te to a ppr op ri at e sp eed 4. 2 Ad ju st la ne po si ti on Lo w/ Fl ex ible Me d. In cr ea si ng to Hi gh Hi gh Lo w Vi su al De ma nd s & In fo So ur ce s • Pr im ar ily en vi ro nm ent dr iv en • Cu rv at ur e pe rc ep ti on cu es • Ob se rv in g ro ad wa y co nd it io ns • Mo st fixa ti on s to ta ngent po in t • Ve hi cl e po si ti on in fo rm at io n E ffe ct iv e In fo Mo de s • Ad vi so ry /m e ssa ge si gn s • No n- ve rb al (e .g ., ch ev ro ns ) an d di re ct in fo (e .g ., de lin ea to rs ) • Di re ct in fo on ly (l an e ma rk in gs ; ra is ed ma rk er s) • No co ns tr ai nt s Ve hi cl e- Co nt ro l De ma nd s • No ne • An ti ci pa to ry po si ti on in g • Cu rv e cu tti ng • Co nt in uous he ad in g ad ju st me nt s • La ne po si ti on ad ju st me nt s Pr im ar y Sp eed In fl uenc es • Pr ev io us ro ad wa y el em ents & si gn ag e • Ex pe ct at io ns & cu rv at ur e cu es • Ex pe ct at io ns & la te ra l a cce le ra ti on • Po st ed sp eed or ex pe ct at io ns

Discussion The information about driving tasks in the previous page is taken from the task analysis described in Tutorial 3 that breaks down curve driving into its perceptual, cognitive, and psychomotor components. A key concept for understanding the curve driving task is the visual and vehicle-control demand, which refers to the amount of time that drivers are required to focus their attention on curve driving activities, such as acquisition of visual information and maintaining vehicle control, to the exclusion of other activities they could otherwise be doing while driving (e.g., scanning for hazards, viewing scenery, changing the radio station, etc.). Visual demands: During the Approach segment, the time and effort that drivers typically spend acquiring information needed to safely navigate a curve is low and driven primarily by the driving environment (e.g., other vehicles, scenery). During Curve Discovery, visual demands increase to high levels at the point of curvature, as drivers scan the curve for information that they need to judge the degree of curvature. Visual demands are highest just after the point of curvature (Entry and Negotiation segment) and drivers spend most of their time looking at the tangent point to keep their vehicle aligned with the roadway (1, 2, 3). For more gradual curves (e.g., 3 degrees), drivers spend more time looking toward the forward horizon than the tangent point (3). Vehicle-control demands: The driver workload imposed by the need to keep the vehicle safely within the lane is minimal up through the end of the Curve Discovery segment, at which point many drivers will adjust their lane position to facilitate curve cutting. Demands are highest during the Entry and Negotiation segment as drivers must continuously adjust the vehicle trajectory to stay within the lane. Moreover, these demands are higher for curves with a shorter radii and smaller lane width (1). During the Exit segment, drivers may adjust their lane position with minimal time pressure, unless there is another curve ahead. Effective information modes: The type of curve-related sign/delineator information that is most likely to be useful to drivers differs in each curve segment. During the Approach, drivers have fewer visual demands and have more time available to read more complex signs, such as speed advisory signs. During the Curve Discovery segment, conspicuous non-verbal information, such as chevrons, are more effective because drivers spend more time examining the curve and have less time available to read, comprehend, and act on text-based information. During Entry and Negotiation, drivers spend most of their time looking at the tangent point, and only direct information presented where they are looking (e.g., lane markings) or information that can be seen using peripheral vision (e.g., raised reflective marking at night) should be relied upon to communicate curve information. Speed selection: Driver expectancy and speed-advisory sign information form the primary basis for speed selection; however, the effectiveness of advisory information may be undermined by expectancy and roadway cues (4). Curve perception also plays an important role in speed selection and inappropriate curvature judgments (e.g., in horizontal curves with vertical sag). Once drivers are in the curve, lateral acceleration felt by drivers and likely vehicle handling workload provide the primary cues for adjusting speed. Expectancy effects: Driver expectations about a curve and, more broadly, design consistency are important factors in drivers’ judgments about curvature and corresponding speed selection during the Curve Discovery segment (1). While direct cues, such as lane width and the visual image of the curve, influence speed selection, expectations based on previous experience with the curve and roadway (e.g., previous tangent length) also significantly influence speed selection (4). Mitigations to recalibrate driver expectancies (e.g., via signage) would likely be most effective prior to the Curve Discovery segment. Design Issues Visual demands appear to be related linearly and inversely to curve radius, but not to deflection angle. Curves sharper than 9 degrees are significantly more demanding than shallower curves or tangents, however, there is no clear, unambiguous threshold regarding what constitutes a sharp curve based on workload data (1, 2). Also, curve direction does not seem to affect workload (2). Additionally, it is unclear whether the 75 to 100 m length of the Curve Discovery segment is based on distance or time. The primary studies that investigated visual demand used the same fixed 45 mi/h travel speed, so it is currently unknown whether the 75 to 100 m fore-distance applies with other speeds (1, 2). Cross References The Influence of Perceptual Factors on Curve Driving, 6-4 Speed Selection on Horizontal Curves, 6-6 Countermeasures for Improving Steering and Vehicle Control Through Curves, 6-8 Countermeasures to Improve Pavement Delineation, 6-10 Signs on Horizontal Curves, 6-12 Key References 1. Krammes, R. A., Brackett, R. Q., Shafer, M. A., Ottesen, J. L., Anderson, I. B., Fink, K. L., Collins, K. M., Pendleton, O.J., and Messer, C.J. (1995). Horizontal Alignment Design Consistency for Rural Two-Lane Highways. Final Report. (Report FHWA-RD-94-034). McLean, VA: FHWA. 2. Fitzpatrick, K., Wooldridge, M. D., Tsimhoni, O., Collins, J. M., Green, P., Bauer, K. M., Parma, K. D., Koppa, R., Harwood, D. W. Anderson, I., Krammes, R. A., and Poggioli, B. (2000). Alternative Design Consistency Rating Methods for Two-Lane Rural Highways. Final Report. (Report FHWA-RD-99-172). McLean, VA: FHWA. 3. Serafin, C. (1994). Driver Eye Fixations on Rural Roads: Insight into Safe Driving Behavior. Interim Report. (Report No. UMTRI-94-21). Ann Arbor: University of Michigan Transportation Research Institute. 4. Fitzpatrick, K., Carlson, P., Brewer, M. A., Wooldridge, M. D., and Miaou, S.-P. (2003). NCHRP Report 504: Design Speed, Operating Speed, and Posted Speed Practices. Washington, DC: Transportation Research Board. 6-3 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0

Ac ce pta ble Ra ng e Ac ce pta ble Ra ng e HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 THE INFLUENCE OF PERCEPTUAL FACTORS ON CURVE DRIVING Introduction The perceptual factors in curve driving refer to the driver’s use of visual information to assess the curvature of an upcoming curve. This activity is important because a driver’s perception of an upcoming curve’s radius forms the primary basis for making speed and path adjustments prior to curve entry. The curve radius as seen from the driver’s perspective is called the apparent radius. Although drivers will use speed information from signs, in practice, driver speed selection in curves is heavily influenced by roadway features (1), and the apparent radius appears be the primary determining factor of speed at curve entry (2). The primary design challenge regarding curve perception is that the apparent radius can appear distorted—either flatter or sharper—depending on the topography and other road elements. Of particular concern are combination curves that include a vertical sag superimposed on a horizontal curve. From the driver’s perspective, this combination makes the horizontal curve appear flatter than it actually is (See A in the figure below). Consequently, drivers may be inclined to adopt a curve entry speed that is faster than appropriate based on horizontal curvature alone. Design Guidelines Sag horizontal curves that have a visual appearance (apparent horizontal radius) that is substantially different from the plan radius should be given careful consideration because they may lead to curve entry speeds that are faster than expected based on horizontal curvature alone. Based Primarily on Based Equally on Expert Judgment Based Primarily on Expert Judgment and Empirical Data Empirical Data A B From Long Tangent From Preceding Curve Apparent Curvature 9000 9000 Actual Curvature Ideal trajectory Trajectory along apparent curvature Curve entry speed based on R ad iu s of S ag V er tic al C ur va tu re ( m ) R ad iu s of S ag V er tic al C ur va tu re ( m ) Potential encroachment Recommended8000 8000 Range 7000 7000Recommended Range 6000 6000 5000 5000 4000 4000 3000 2000 Unacceptable Range 1000 3000 2000 Unacceptable Range 1000 apparent curvature is too high to safely traverse the actual curvature 0 0 0 10 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 0 10 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 Radius of Horizontal Curve (m) Radius of Horizontal Curve (m) A. A vertical sag curve produces a visual image (shaded roadway) that a driver would perceive as having an apparent radius that is larger than the actual radius. B. Nomographs indicating vertical and horizontal curve radius combinations that result in apparent radii that may result in curve entry speeds that are unintentionally faster than expected based on horizontal curvature alone (red shaded region), and which possibly represent a safety risk (2). Note that the nomographs present vertical curvature in terms of radius (in meters) and not K, which is the typical approach for representing vertical curvature. The reason for presenting curvature as a radius is that the geometric calculations for computing visual distortion rely on circular arcs. The nomographs can be used to provide a “rule of thumb” check for potentially problematic curve combinations assuming the vertical curvature component can be generally approximated by a circle with an arc intersecting the low point of Type III curves and vertical points of curvature on both sides. 6-4

Discussion Curve perception is an im portant part of curve driving because, in the absence of extensive experience with a curve, drivers mu st rely on their judgm ents about a curve to select a safe speed for curve entry. Speed signage inform ation can assist drivers; however, evidence suggests that this inform ation is not a prim ary source for speed selection in curves ( 1 ). Therefore, driver expectations (influenced by design consistency) and the visual inform ation the driver obtains about the curve are the prim ary basis for speed selection. Sag horizontal curves can cause drivers to significantly underestim ate the sharpness of a curve because of a visual distortion from the driver’s viewing perspective; i.e., the apparent radius appears to be longer than the plan radius. Thus, these sag horizontal curves, are also associated with higher entry speeds and crash rates ( 2 , 3 ). The optical aspects of this phenom enon have been derived analytically, and the results were used to ma ke the nom ographs presented on the previous page. Horizontal and vertical curve radius co mb inations that fall in the unacceptable range are associated with significant visual distortion, and also associated with higher than 85 th percentile speeds and higher crash rates ( 2 ). Note that this validation is based on European data, and these findings have not been investigated on US roads. However, the optical properties of this phenom enon are universal and should be equally applicable to all drivers ( 4 ). This analytical work also assum es a 75 m viewing distance, which is comparable to the start of the Curve Discovery segm ent of curve driving, in which drivers spend mo st of their time inspecting the curve. Distortion effects may be reduced som ewhat at further viewing distances; however, assum ing a 75 m viewing distance is consistent with driver behavior and is more conservative. Visual distortion also occurs when crest vertical curves are superim posed on horizontal curves; such curves appear sharper than the plan radius. This typically results in slower 85 th percentile entry speeds ( 2 , 3 ). However, a crest horizontal curve with a vertical curvature that approxi mate s a circular radius of less than 3 times the horizontal curve radius could present a discontinuous visual im age of the curve (e.g., the part of the roadway just behind the crest is occluded) ( 2 ). Such a crest horizontal curve is potentially inconsistent with driver expectations and could com prom ise roadway safety by causing drivers to suddenly brake hard if they are surprised by the curve appearance. However, there are currently no em pirical data showing that this is an actual safety issue. Design Issues A su mma ry of the relevant research findings regarding curve perception in general and the corresponding degree of em pirical support is shown in the table below. While no specific values or reco mme ndations can be ma de for these aspects, it is useful to take them into consideration during curve design, especially if other aspects of the curve design suggest that there may be a potential problem with driver perception of the curve radius. Aspect Effect Empirical Support Superimposed Vertical Sag Makes a curve appear flatter Strong Cross Slope For sag horizontal curves, the greater the cross slope and lane width, the greater the apparent flattening of the horizontal curve Analy tical evidence Superimposed Vertical Crest Makes a curve appear sharper and ma y cause discontinuities in curve Strong Deflection Angle Holding radius constant, greater deflection angle makes the curve appear sharper, especially for smaller radii Moderate Delineators Delineators provide drivers with more information to judge the curve radius, which improves accuracy of these judgments Moderate Spiral Ma y make curve appear flatter, or make curve perception more difficult, because the onset of the curve is less apparent Indirect Signage Drivers perceive curve as “riskier” if signs indicate that the curve is hazardous Suggestive Cross References Task Analy sis of Curve Driving, 6-2 Key References 1. Fitzpatrick, K., Carlson, P., Brewer, M. A., Wooldridge, M. D., and Miaou, S.-P. (2003). NCHRP Report 504: Design Speed, Operating Speed, and Posted Speed Practices . Washington, DC: Transportation Research Board. 2. Appelt, V. (2000). New approaches to the assessm ent of the spatial alignm ent of rural roads—apparent radii and visual distor tion. Proceedings of the 2nd International Symposium on Highway Geometric Design (pp. 620-631). Cologne, Germ any: Verlag. 3. Hassan, Y., and Easa, S. M. (2003). Effect of vertical alignm ent on driver perception of horizontal curves. Journal of Transportation Engineering, 129 (4), 399-407. 4. Bidulka, S., Sayed, T., and Hassan, Y. (2002). Influence of vertical alignment on horizontal curve perception: Phase I: Examining the hypothesis. Transportation Research Record, 1796 , 12-23. 6-5 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0

SPEED SELECTION ON HORIZONTAL CURVES Introduction Various sources attem pt to exam ine speed data for roadway geom etry and to deter mi ne desirable speeds for horizontal curves. AASHTO policy defines design speed as “a selected speed used to de termine the various geometric design features of the roadway” ( 1 ). The design speeds on horizontal curves should be set at a value deter min ed by AASHTO policy and factors deter min ed from a survey of state DOTs. AASHTO policy ( 1 ) considers factors such as functional classification, rural vs. urban environm ent, and terrain type; state DOTs typically consider factors such as functional classification, legal speed limit (as well as legal speed limit plus an adjustment value of 5 or 10 mi/h), anticipated volume, terrain type, development, costs, and design consistency. Design Guidelines A number of vehicle, driver, and roadway variables should be considered when determin ing speed limits for horizontal curves. A procedure to calculate appropriate speeds has been adapted from Charlton and de Pont ( 2 ) and is outlined below. If these factors are common at an intersection location, then consideration s hould be given to modify ing the gap acceptance design assumptions. Step Procedures for Determining Curve Advisory Speed Limits 1 Determine curve radius ( R ), superelevation, and offset distance from center of lane to any visual obstruction ( O ). 2 Determine the vehicle’s maximum possible lateral acceleration and braking coefficient. The maximum lateral acceleration is limited by rollover stability for most heavy vehicles and by tire adhesion for passenger cars. Typical values to use for dry conditions are 0.35 g for laden heavy vehicles, 0.7 g for buses and SUVs, and 0.8 g for passenger cars. The braking coefficient reflects the maximum braking efficiency that can be achieved and should be 0.9–1.0 for passenger cars and 0.5–0.6 for heavy vehicles. Assume a reaction time (Tr) of 2 s. 3 Calculate the maximum possible speed (in km/h) limited by lateral acceleration using the formula: ) ( 127 tio n supereleva c lateral_ac R V 4 4.1 From this speed, calculate the safety factor (SF) using the equation: 2 00004762 . 0 03476 . 0 1 V V SF 4.2 Divide the maximum lateral acceleration value by the safety factor ( SF ), and recalculate the speed using the equation in step 3. This is the desirable maximum speed limited by lateral acceleration, V acc . tion supereleva SF c lateral_ac R V acc 127 5 5.1 Calculate the sight distance using the equation: R O R R SD acc 1 cos 2 5.2 Based on a safety factor of 2, set the braking coefficient ( d ) to half the maximum braking efficiency value. Then, set the stopping sight distance equal to the si ght distance calculated above and solve for speed ( V sight ) in the following stopping distance equation: d SD T T d V d V V T SD SD acc r r sight sight sight r acc stop 254 4 6 . 3 6 . 3 127 254 6 . 3 2 2 6 The maximum desirable speed for the particular vehicle in the curve is the lesser of the two maximum speed values, V acc and V sight . Variables V = Vehicle Speed (km/h) R = Curve Radius (m) SD sto p = Stopping Sight Distance O = Offset Distance from center of the lane to the obstruction (m) SD acc = Sight Distance T r = Driver Reaction time (seconds) V acc = Desirable maximum speed limited by lateral acceleration (km/h) d = Braking Coefficient V sight = Desirable maximum speed limited by sight distance (km/h ) Based Primarily on Expert Judgment Based Equally on Expert Judgment and Empirical Data Based Primarily on Empirical Data HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 6-6

Discussion Drivers’ failure to accurately judge the appropriate driving speed on horizontal curves can have safety consequences. The Fatality Analysis Reporting System (FARS) indicates that 42,815 people were killed in 38,309 fatal crashes on the US highway system in 2002. Approximately 25% of these crashes occurred along horizontal curves. These crashes occurred predom inantly on two-lane rural highways that are often not part of the state DOT system. Approximately 76% of curve-related fatal crashes were single-vehicle crashes in which the vehicle left the roadway and struck a fixed object or overturned; conversely only 11% of curve-related crashes were head-on crashes. Speed selection by drivers on horizontal curves reflects a variety of vehicle, driver, and roadway factors. For exam ple, drivers of vehicles with larger engines, and greater acceleration capacity, approach curves differently than other drivers ( 3 ). Experienced and mi ddle-aged drivers report less accurate estim ates of perceived speed than do younger and less- experienced drivers along roadway curves ( 4 ). Visual mi sperceptions ma y occur when the horizontal curve is com bined with a vertical curve. For exam ple, on-road records of vehicle speed were dem onstrated to be consistent with a mi sperception hypothesis on crest com binations ( 5 ); i.e., the horizontal radius is perceived to be shorter than it actually is. In a safety research study ( 6 ), relationships of safety to geom etric design consistency m easures were found to predict speed reduction by mo torists on a horizontal curve relative to preceding curve or tangent, average radius, and rate of vertical curvature on a roadway section and ratio of an individual curve radius to the average radius for the roadway sections as a whole. A review of vehicle speed distributions and the variation of vehicle speed around single road curves found that the pattern of variation in vehicle speeds along a road curve was highly dependant on the level of curvature; this effect was mo re pronounced for curves of radius less than 250 m ( 7 ). While radius of curvature is not the only factor that influences selected speed on horizontal curves ( 8 ), it may be the most important factor ( 9 ). Deter mi ning speeds for horizontal alignm ent is a com plex mi x of personal judgm ent, em pirical analysis, and AASHTO/state DOT guidelines. A num ber of sources provide equations and procedures that reflect the com plexity of speed selection on curves by drivers. A series of speed prediction equations for passenger vehicles on two-lane highways as a function of various characteristics of the horizontal curve is provided in Anderson, Bauer, Harwood, and Fitzpatrick ( 6 ). A series of steps that can be used to deter mi ne ma xi mu m desirable speed is provided in Charlton and de Pont ( 2 ). Design Issues Transportation Research Circular 414 ( 10 ) stated factors contributing to higher crash frequency on horizontal curves include higher traffic volumes, sharper curvature, greater central angle, lack of a transition curve, a narrower roadway, mo re hazardous roadway conditions, less stopping distance, steep grade on curve, long distance since last curve, lower pavem ent friction, and lack of proper signs and delineation. Cross References The Influence of Perceptual Factors on Curve Driving, 6-4 Key References 1. AASHTO (2011). A Policy on Geometric Design of Highways and Streets. Washington, DC. 2. Charlton, S.G., and de Pont, J.J. (2007). Curve Speed Management (Research Report 323). Waterloo Quay, Wellington: Land Transport New Zealand. 3. Bald, S. (1987). Investigation of Determinants of Choice of Speed: Evaluation of Speed Profiles on Country Roads [Abstract]. Darmstadt, Germany: Technical University of Darmstadt. 4. Milosevic, S., and Milic, J. (1990). Speed perception in road curves. Journal of Safety Research, 21(1), 19-23. 5. Bella, F. (2006). Effect of driver perception of combined curves on speed and lateral placement. Transportation Research Board 85th Annual Meeting Compendium of Papers [CD-ROM]. 6. Anderson, I., Bauer, L., Harwood, D., and Fitzpatrick, K. (1999). Relationship to safety of geometric design consistency measures for rural two-lane highways. Transportation Research Record, 1658, 43-51. 7. Mintsis, G. (1998). Speed distribution on road curves. Traffic Engineering and Control, 29(1), 21-27. 8. Andjus, V., and Maletin, M. (1998). Speeds of cars on horizontal curves. Transportation Research Record, 1612, 42-47. 9. Bird, R. N., and Hashim, I. H. (2005). Operating speed and geometry relationships for rural single carriageways in the UK. Proceedings of the 3rd Symposium on Highway Geometric Design [CD-ROM]. Washington, DC: Transportation Research Board. 10. Transportation Research Board, National Research Council (1993). Transportation Research Circular 414: Human Factors Research in Highway Safety. Washington, DC. 6-7 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0

C OUNTERMEASURES FOR I MPROVING S TEERING AND V EHICLE C ONTROL T HROUGH C URVES Introduction Successful navigation of curves depends on accurate steering and speed cont rol in order to mi ni mi ze lateral acceleration within the lane. Design of alignm ents that c onform to driver expectations and typical behaviors will enhance the driver’s ability to control the vehicle. This guideline provides strategies for im plem enting curve geom etries that help drivers main tain proper lane position, speed, and lateral control through curves. Delineation treatments that improve vehicle control are presented in the “Countermeasures to Improve Pavement Delineation” guideline. The following guidelines present strategies for designing geometric features that will enhance steering control. Curvature Minimize the use of controlling curvature (i.e., maximum allowable curvature for a given design speed). Spirals Spiral transition curves should be used whenever possible, particularly for curves on roads with high design speeds (e.g., 60 m i/h or greater). Spiral curve lengths should equal the distance traveled during steering tim e (i.e., 2 to 3 s depending on radius). The reco mme nded curve radius for two-lane highways with a speed lim it of 50 mi/h is 120 to 230 m, with clothoid param eters between 0.33 and 0.5 R. Reverse Curves Do not use tangent sections in reverse curves when the distance between the exit of the first curve and the entrance of the second curve is short enough to encourage a curved path through the tangent (e.g., 80 m or less for two-lane highways and 135 m for freeways). Superelevation Superelevation should be designed to result in zero lateral acceleration through the curve at design speed. Design Consistency Avoid sharp, isolated curves and maintain consistency in the design of superelevation, road width, and other curve features to improve conformance with drivers’ expectations. The figure below illustrates the various concepts that describe how drivers navigate a curve: visual components related to guidance and lane-keeping, the path choice model, and the combination of processes that govern curve traversal. • Dr Ideal Trajectory Actual Trajectory Direction of gazeHeading Far region: curvature information (anticipatory process) Near region: position-in-lane information (compensatory process) T = Tangent point ractual = Vehicle radius of travel rideal = Ideal radius of travel dmin = Minimum acceptable distance from lane edge iver enters curve to the left of the lane center • Far region provides cues for predicting curvature and steering angle in closed- loop anticipatory control process. • Near region (≤7 degrees down from horizon) provides cues for correcting deviations from path in open-loop compensatory control process. • Driver follows trajectory with radius of curvature (ractual) greater than radius at center of lane (rideal) and that brings the vehicle to a minimum distance (dmin) from the roadway edge line at its apex. • Driver fixates on curve tangent point through the curve. Adapted from Donges (1); Levison, Bittner, Robbins, and Campbell (2); and Spacek (3). Figure not to scale. dmin ractualrideal T Based Primarily on Expert Judgment Based Equally on Expert Judgment and Empirical Data Based Primarily on Empirical Data Design Guidelines HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 6-8

Discussion The steering control task has been m odeled as a two-level process com posed of an open-loop anticipatory com ponent (far view) for predicting curvature and steering angle, and a closed-loop com pensatory com ponent (near view) for correcting deviations from the desired path ( 1 ). However, this two-level m odel does not adequately describe som e path-decision behaviors such as curve-cutting. Also, drivers often ma ke anticipatory steering actions based on an internal estim ate of the vehicle characteristics and on previously perceived curvature, rather than on direct visual feedback, while paying attention to other aspects of the driving task ( 4 ). Geom etric alignm ent and delineation features affect the driver’s perception of curvature and therefore influence curve entry speed. Curve geom etries that do not m eet the driver’s perceptual expectations may result in inappropriate entry speeds that require speed and steering corrections within the curve in order to avoid excessive lateral acceleration and a potential loss of control. Inaccuracies in anticipatory assessm ent prior to curve entry generally increase with curvature, and com pensatory control actions to correct these errors are greatest in sharp curves ( 4 , 5 ). In general, drivers tend to cut curves. In one study ( 3 ), al mo st one-third of drivers cut left-hand curves and 22% cut right-hand curves. Drivers com pensate for inadequate steering adjustm ent at curve entry by following a trajectory with a radius that is larger than the ideal radius (i.e., radius at the center of the lane), with the vehicle traveling within some mi ni mu m distance of the edge line at its apex ( 2 , 7 ). Vehicle path radius at the point of highest lateral acceleration correlates with higher crash rates. Design Issues Curvature: Road curvature significantly affects average lateral position error. As curves becom e sharper, there is a corresponding increase in workload, which can result in an increase in edge line encroachm ents on the inside lane ( 6 , 7 ). Restrictive geom etric characteristics (e.g., sharper curves, narrower shoulders, and steeper grades) are mo re likely to lead to centerline encroachments than those that are less constraining; however, high curvature has the greatest adverse effect on crash rates and driving perform ance in horizontal curves. Spiral curves: Spirals that are designed to matc h drivers’ natural steering behavior offer a gradual increase in centrifugal force and facilitate superelevation transitions, which can improve the vehicle’s lateral stability ( 6 , 7 , 8 ). However, overly long spiral transitions can lead to mi sleading perception of the sharpness of curvature, inappropriate entry speed, and unexpected steering and speed corrections within the curve. The mo st desirable spiral length is equal to the distance traveled during the steering time (nominally 2 to 3 s depending on radius). Reverse curves: Tangent sections of appropriate length can provide effective transitions between curves in a reverse curve alignm ent. However, if the tangent section is too short, drivers ma y follow a curved rather than straight trajectory through the tangent section ( 7 ). To matc h the alignm ent to drivers’ typical steering behavior, the transitional tangent should be long enough to allow straightening of the vehicle through the transition (if possible); otherwise, the transitional tangent should not be used. Design consistency: Drivers are more likely to make appropriate sp eed and steering decisions when the roadway design m eets their perceptual expectations. Consistency in curve features, such as superelevation, lane width, curvature, etc., help reduce workload and therefore im prove stability in steering control ( 6 ). Cross References The Influence of Perceptual Factors on Curve Driving, 6-4 Speed Selection on Horizontal Curves, 6-6 Counterm easures to Im prove Pavem ent Delineation, 6-10 Key References 1. Donges, E. (1978). Two-level m odel of driver steering behavior. Human Factors, 20 (6), 691-707. 2. Levison, W. H., Bittner, A. C., Robbins, T., and Cam pbell, J. L. (2001). Development of Prototype Driver Models for Highway Design. Task C: Develop and Test Prototype Driver Performance Module (D PM). Washington, DC: FHWA. 3. Spacek, P. (2005). Track behavior in curve areas: Attem pt at typology. Journal of Transportation Engineering, 131 (9), 669-676. 4. Godthelp, H. (1986). Vehicle control during curve driving. Human Factors, 28 (2), 211-221. 5. Si ms ek, O., Bittner, A. C., Levison, W. H., and Garness, S. (2000). Development of Prototype Driver Models for Highway Design. Task B: Curve-Entry Speed-Decision Vehicle Experiment . Washington, DC: FHWA. 6. Reinfurt, D. W., Zegeer, C. V., Shelton, B. J., and Neum an, T. R. (1991). Analysis of vehicle operations on horizontal curve s. Transportation Research Record, 1318 , 43-50. 7. Said, D. G., Hassan, Y., and Abd El Halim , O. (2007). Quantif ication and utilization of driver path in im proving design of h ighway horizontal curves. Transportation Research Board 86th Annual Meeting Compendium of Papers [CD-ROM]. 8. Perco, P. (2006). Desirable length of spiral curves for two-lane rural roads. Transportation Research Record, 1961 , 1-8. 6-9 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0

COUNTERMEASURES TO IMPROVE PAVEMENT DELINEATION Introduction This guideline describes countermeasures that support improvements in curve detection and driver performance through the use of pavement surface markings, such as edge lines, raised retroreflective pavement markers (RRPM), transverse stripes, etc. These markings provide primarily non-verbal cues that promote improved vehicle control through earlier detection and recognition of curves, reductions in speed, and adjustments to lateral position. Design Guidelines General Use surface delineations that are characterized by small gaps, long dashes, and short repetition cycles. Use combinations of treatments wherever practical to increase overall effectiveness. Edge line/ Centerline Use edge lines when curves are sharp or frequent, on narrow roads, or in the vicinity of crossing roadways or major driveways. Use the widest possible edge lines and centerlines to maximize visible surface area. When possible, use striping materials with highly retroreflective characteristics to implement edge lines and centerlines. RRPM Combine RRPM with edge lines/centerlines. Use pairs of RRPM on the outside edges of the centerline for very sharp curves ( 12 degrees); for flatter curves, single RRPMs are sufficient. Place RRPMs 244 m in advance of the curve. Space markers at 40 m intervals for sharp curves and 80 m intervals for flatter curves. Transverse Stripes When practical, implement transverse stripes as graduated rumble strips. Space stripes to achieve 0.5 s intervals at the desired deceleration rate (e.g., 0.9 m/s2) “SLOW” text with arrow Use “SLOW” with arrow surface markings in the tangent section approximately 70 m before the curve to augment treatments in high-hazard areas or at sharp curves. The following table indicates various pavement marking treatments and their strengths for enhancing speed reduction, lane-keeping, and curve detection and recognition. Treatment Type Strengths General – Surface markings Strongest curvature cues and short-range steering control (compensatory control) General – Post-mounted chevrons Strongest guidance cues and long-range guidance (anticipatory control) Treatment Combinations Superior effectiveness compared with individual treatments Edge line/Centerline Strongest for curve recognition, curvature perception, and reduction of lateral variability. Discontinuities in edge line aid in recognizing upcoming intersections, driveways, etc. RRPM Improving visibility of edge lines and centerlines. Reducing lane encroachments. Both visual and rumble effects provide encroachments cues. Transverse Stripes Speed reduction. May be more effective at reducing higher (> 85th percentile) speed driving than lower speed driving. “SLOW” Text with Arrow Speed reduction and curve ahead warning. Based Primarily on Expert Judgment Based Equally on Expert Judgment and Empirical Data Based Primarily on Empirical Data HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 6-10

Discussion Road delineations provide cues that assist drivers in detecting curves and assessing the level of curvature. Road surface markings provide the strongest curvature cues and are best for providing short-range steering control cues (compensatory control—see “Countermeasures for Improving Steering and Vehicle Control Through Curves”), while chevron designs on post-mounted panels give the strongest guidance cues and are best for long-range guidance (anticipatory control). Under conditions of reduced visibility, steering performance improves in the presence of road surface delineations that are characterized by small gaps, long dashes, and short repetition cycles. Edge lines improve perception of curvature, curve recognition distance, and lane-position stability. Roads with edge lines exhibit fewer crashes than those without edge lines, particularly in combination with narrow widths, wet pavement, and/or high-hazard areas (1). Surface area has the greatest effect on edge line (and centerline) visibility— effectiveness increases with wider edge lines. Also, the effectiveness of these stripes increases with the level of retroreflectivity. Raised reflective pavement markers are highly effective at improving curve visibility and reducing crashes, especially when used in combination with centerlines and edge lines (2). They can be particularly useful as a cue for warning of lane encroachment because the raised marker provides tactile as well as visual stimulus. As with edge lines, the effectiveness of RRPMs increases with retroreflectivity. Transverse stripes refers to painted or taped stripes that are applied perpendicularly across the roadway alignment. Typically, these stripes are separated by decreasingly graduated spacings in order to encourage speed reduction by creating a sensation of increased speed when the vehicle is traveling at constant speed. The effectiveness of transverse stripes has been mixed; while some studies report reductions in speed at curve entry (3), others report either no reduction or a slight increase in speed (4). Transverse stripes are most effective when implemented as rumble strips because they provide both visual and tactile stimuli. “Slow” text with arrow refers to the word “Slow” marked in elongated letters with an arrow above it pointing in the direction of the curve and transverse lines before and after the symbols. This treatment may be effective at speed reduction, especially in late night driving when drivers are more likely to be impaired by fatigue or alcohol (5). Combinations of treatments are generally more effective than any single treatment, especially when the combination includes rumble strips. Curve recognition, lane position, and number of encroachments are improved when RRPMs are used in conjunction with edge line/centerline markings compared with single treatments. Design Issues In general, centerline treatments tend to cause drivers to shift lateral position away from the centerline, while edge line treatments result in a lateral shift toward the centerline. RRPMs may reduce nighttime corner cutting in left-hand curves but increase corner cutting in right-hand curves (6). Several treatments, such as transverse stripes and widening of inside edge markings at the curve, may have a greater effect on driver performance for high-speed drivers (above 85th percentile speeds) than for lower-speed drivers. These treatments should be considered in hazard areas where speed is a prevalent factor in elevated crash rates (3). Cross References Speed Selection on Horizontal Curves, 6-6 Countermeasures for Improving Steering and Vehicle Control Through Curves, 6-8 Key References 1. Tsyganov, A. R., Machemehl, R. B., and Warrenchuk, N. M. (2005). Safety Impact of Edge Lines on Rural Two-Lane Highways. (CTR Research Report 0-5090-1). Austin: University of Texas Center for Transportation Research. 2. Nemeth, Z. A., Rockwell, T. H., and Smith, G. L. (1986). Recommended Delineation Treatments at Selected Situations on Rural State Highways. Part 1 (Revision). (Report EES-627-PT-1, FHWA/OH-86/009). Columbus: Ohio State University, College of Engineering. 3. Vest, A., and Stamatiadis, N. (2005). Use of warning signs and markings to reduce speeds on curves. Proceedings of the 3rd International Symposium on Highway Geometric Design [CD ROM]. Washington, DC: Transportation Research Board. 4. Macaulay, J., Gunatillake, T., Tziotis, M., Fildes, B., Corben, B., and Newstead, S. (2004). On-Road Evaluation of Perceptual Countermeasures. (Report CR 219) Civic Square, ACT: Australian Transport Safety Bureau. 5. Retting, R. A., and Farmer, C. M. (1998). Use of pavement markings to reduce excessive traffic speeds on hazardous curves. ITE Journal, 68(9), 6. 6. Zador, P., Stein, H. S., Wright, P., and Hall, J. (1987). Effects of chevrons, post-mounted delineators, and raised pavement markers on driver behavior at roadway curves. Transportation Research Record, 1114, 1-10. 6-11 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0

S IGNS ON H ORIZ ONTAL C URVES Introduction Prior to a change in the horizontal alignm ent of a roadway, inform ation about this change should be conveyed to drivers via roadway signs. This inform ation should be co mm unicated in a concise and efficient ma nner such that drivers have tim e to process the information and adjust their speed as well as alter the vehicle path appropriately. Notification of an upcom ing curve is typically conveyed using curve warning signs, which indicate whether the curve is to the right or the left; they are som etim es accom panied by advisory speed signs. The use of dynam ic warning signs to alert drivers of a curve and/or their vehicle speed has also gained acceptance as an effective m eans of co mm unication. Researchers disagree as to how advance warnings should be presented to drivers, i.e., through text or through symbols. But all agree that the key to effective warning is to notify the driver of the upcoming curve so that the driver can change the speed or path of the vehicle—or both. Individual studies on the effectiveness of advance warning signs vary considerably with respect to sign placements, sign messages, horizontal curve radii, and driver populations. Designers should consider such variables when making design decisions. Also, any information considered for use in curve signs should not be in conflict with current design standards in publications such as the MUTCD. Design Guidelines The tables below show the guidelines for advance placement of curve warning signs related to advisory /85 th percentile speed, as well as spacing for chevrons—both are presented as a function of posted or advisor y speeds (Adapted from McGee and Hanscom ( 1 )). Advance Placement Distance (ft) for Advisory Speed of the Curve (mi/h) of Posted or 85 th Percentile Speed (mi/h) 10 20 30 40 50 60 70 20 n/ a 1 – – – – – – 25 n/a 1 n/a 1 – – – – – 30 n/a 1 n/a 1 – – – – – 35 n/a 1 n/a 1 n/a 1 – – – – 40 n/a 1 n/a 1 n/a 1 – – – – 45 125 n/a 1 n/a 1 n/a 1 – – – 50 200 150 100 n/a 1 – – – 55 275 225 175 100 n/a 1 – – 60 350 300 250 175 n/a 1 – – 65 425 400 350 275 175 n/ a 1 – 70 525 500 425 350 250 150 – 75 625 600 525 450 350 250 100 1 No suggested distance is provided for these speeds, as the placement location depends on site conditions and other signing to provide an adequate advance warning for the driver. Advisory Speed Limit (mi/h) Chevron Spacing (ft) 15 40 20 80 25 80 30 80 35 120 40 120 45 160 50 160 55 160 60 200 65 200 NOTE: The above spacing distances apply to points within the curve. Approach and departure spacing distances are twice those shown above. Based Primarily on Expert Judgment Based Equally on Expert Judgment and Empirical Data Based Primarily on Empirical Data HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0 6-12

Discussion Numerous studies have shown the effectiveness of advanced warning signs for curves (2, 3, 4, 5). Typical improvements in driving performance are reductions in speed, fewer lane excursions, and generally fewer crashes—see also the table below. From a driver’s perspective, the key advantage of advance warning signs is a notification that a (possibly) unexpected change in the horizontal alignment of the roadway is imminent. Signing can be used to notify the driver of an upcoming curve in many ways, including proper positioning along a driver’s line of sight, fluorescent illumination, flashing beacons (5), or dynamic warnings. In this regard, designers are cautioned to avoid overloading the driver with extraneous information that might distract him or her from the primary task of maintaining safe control of the vehicle (6). Improvement Reference Findings Fluorescent Yellow Microprismatic Chevron Treatments 2 Weighted average decrease in speeds at the curve point of curvature of about 1 mi/h for both the mean and 85th percentile versus the existing standard yellow ASTM Type III signs. 38% overall reduction in edge line encroachments. Fluorescent Yellow Chevron Posts 2 Speeds reduced slightly. Fluorescent Yellow Microprismatic Curve Warning Signs 2 The overall number of vehicles initiating deceleration before reaching the curve warning sign was increased by 20%. However, the study found small and inconsistent effects on speeds approaching curves. Standard Red Reflectorized Border on Speed Limit Sign 2 The red border had the greatest effect on speeds during the day for both passenger vehicles and heavy trucks. Daytime mean and 85th percentile speeds of heavy trucks were found to decrease by 4 mi/h. Addition of Flags, Flashers on Existing Warning Signs 3 The changes made to roadway surface included more reflective centerlines (CLs), more reflective edge lines (ELs), wider ELs, the additional of raised retroreflective pavement markers, and the inclusion of horizontal signing warning of approaching curves. Dynamic Advance Curve Warning System 4 Results found decreases in mean speeds from 2 to 3 mi/h. Different Pavement Markings and Raised Retroreflective Pavement Markers 5 Nighttime average speed reductions for the warning sign with flashing lights (5.1%), the combination horizontal alignment/advisory speed sign (6.8%), and flashing lights on both warning signs (7.5%). Design Issues In a literature synthesis of the knowledge and practice, the physical and performance characteristics of heavy vehicles that interact with highway geometric design criteria and devices were examined (7). The synthesis notes that dynamic curve warning systems for trucks—especially highly accurate, sophisticated systems that incorporate vehicle parameters such as speed and weight—may help warn drivers of curves ahead and mitigate rollover crashes. Cross References Speed Selection on Horizontal Curves, 6-6 Key References 1. McGee, H. W., and Hanscom, F. R. (2006). Low-Cost Treatments for Horizontal Curve Safety. (FHWA-SA-07-002). Washington, DC: FHWA. 2. Gates, T.J., Carlson, P.J., and Hawkins, H.G., Jr. (2004). Field evaluations of warning and regulatory signs with enhanced conspicuity properties. Transportation Research Record, 1862, 64-76. 3. Molino, J., Donnell, E. T., and Opiela, K.S. (2006). Field ratings of nighttime delineation enhancements for curves. Transportation Research Board 85th Annual Meeting Compendium of Papers [CD-ROM]. 4. Monsere, C. M., Nolan, C., Bertini, R. L., Anderson, E. L., and El-Seoud, T. A. (2005). Measuring the impacts of speed reduction technologies: Evaluation of dynamic advanced curve warning system. Transportation Research Record, 1918, 98-107. 5. Vest, A., and Stamatiadis, N. (2005). Use of warning signs and markings to reduce speeds on curves. Proceedings of the 3rd International Symposium on Highway Geometric Design [CD-ROM]. Washington, DC: Transportation Research Board. 6. Zwahlen, H. T. (1987). Advisory speed signs and curve signs and their effect on driver eye scanning and driving performance. Transportation Research Record, 1111, 110-120. 7. Harwood, D. W., Potts, I. B., Torbic, D. J., and Glauz, W. D. (2003). CTBSSP Synthesis of Safety Practice 3: Highway/Heavy Vehicle Interaction. Washington, DC: Transportation Research Board. 6-13 HFG CURVES (HORIZONTAL ALIGNMENT) Version 2.0