Click for next page ( 20


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 19
HSIP funds and all of its HRRR funds to safety and no intersection in the state averages one projects on local systems. fatal crash per year. Minnesota (48% of fatal crashes on the local The average two-lane rural state highway aver- system) has provided a local version of CMAT ages 1.5 crashes per mile per year and 0.01 fatal to all cities and counties. MnCMAT contains crashes per mile per year. 10 years of data, and up to 73 data items are The average county highway averages 0.5 provided for each crash including route, loca- crashes per mile per year and 0.003 fatal tion, date/day/time, severity, vehicle actions, crashes per mile per year. crash causation, weather, road characteristics, The point is that the mature analytical systems and driver condition. Mn/DOT has also pro- that safety professionals are familiar with are primar- vided technical assistance through a series of ily focused on finding locations with unusually high safety workshops around the state and has numbers of crashes, which most often are not the revised its approach to the HSIP. The safety locations where the majority of the severe crashes are fund is disaggregated by district based on the actually occurring. In response to this challenge, state distribution of fatal crashes around the state, and national agencies have been working to identify and within each district the funds are split based at-risk rural locations by developing tools that are not on the distribution of fatal crashes between the based just on crash data but also take into account state and local systems. This new approach has identifying features such as design characteristics and directed more than 60% of HSIP funds to traffic volumes. Examples of these tools are discussed Mn/DOT's rural districts, and almost 50% of below. the safety funds are reserved for projects devel- oped by local agencies for implementation on local roads. SafetyAnalyst In addition to noting the lessons they have learned, This is a new suite of analytical tools for identi- the states indicated some challenges that will have to fying and managing a systemwide program of site- be addressed before inclusion of local road authorities specific improvements to enhance highway safety in the safety planning process becomes routine. These by cost-effective means. The package was devel- challenges are developing methodologies and tools for oped by FHWA and partner state and local agencies. identifying candidate sites for safety investment in The software can be used to identify the frequency rural areas and the lack of safety-related experience and percentage of specific crash types systemwide, among the staff at the local road authorities. on particular segments of a road network, or at indi- vidual high-crash locations (black spots). The pro- gram can also be used to characterize the need for METHODOLOGIES AND TOOLS TO systemwide engineering improvements such as edge SUPPORT SAFETY PLANNING EFFORTS treatments and cable median barriers. A key expected One of the key challenges identified by the par- benefit of SafetyAnalyst is automation of the man- ticipating states is that the analytical processes for ual safety analyses being conducted by some road authorities. identifying candidate sites for safety investments in The SafetyAnalyst package consists of six tools: rural areas (rural intersections and rural highway segments on both the state and local systems) are not Network screening well developed, and the basic processes are not Diagnosis understood by safety engineers and analysts. Most Countermeasure selection previous efforts to refine analytical processes have Economic appraisal focused on improving the statistical methods for Priority ranking identifying high-crash locations. However, most of Countermeasure evaluation the rural locations where most of the severe crashes occur have had few or no crashes during a typical The network screening tool is used to identify 3- to 5-year study period. For example, in Minnesota sites that have the potential for safety improvement based on higher-than-expected crash frequencies. The The average rural intersection averages 0.5 diagnosis tool generates collision diagrams and helps crashes per year and 0.01 fatal crashes per year, the user understand the nature of collision patterns 19

OCR for page 19
that may exist at screened sites. The tool includes a develop their own techniques and tools. The partici- diagnostic expert system that asks the user questions pating states' final comments reflected a nearly uni- about specific sites and specific crash scenarios in versal concern that SafetyAnalyst would not be used order to suggest specific countermeasures, and the any time soon to assist with safety planning on local tool considers both engineering and human factors systems because local agencies would not have the criteria. The countermeasures selection tool is inte- necessary databases documenting roadway and inter- grated with the diagnosis tool and presents users with section features. a suggested set of countermeasures for further consid- eration. The economic appraisal tool is used to assess the economic viability of each of the countermea- United States Road Assessment Program sures, using four economic appraisal methods. The (usRAP) economic appraisal tool also includes an optimization This is a new methodology being developed by algorithm that can consider multiple sites and multi- the AAA Foundation for Traffic Safety (AAAFTS) to ple candidate countermeasures at each site and then evaluate safety improvement opportunities on a road suggest a set of sites and countermeasures that pro- network selected by a highway agency and to identify vides the maximum safety benefit within a user-spec- cost-effective safety improvements. ified budget. The priority ranking tool is integrated The road network to be considered is selected by with the economic appraisal tool and ranks the candi- a participating highway agency in consultation with date treatment sites and countermeasures using a usRAP. Three protocols are included by usRAP: risk range of economic, safety, and project cost measures. mapping, star ratings, and countermeasures selection. The SafetyAnalyst software tools require access Risk maps only require information about severe crash to a database that includes roadway/intersection char- locations and a limited amount of information on road- acteristics, traffic volumes, and crash data for the road way features and traffic volume characteristics. While network to be evaluated. Many of the data elements more reactive in nature, risk mapping provides a sys- required for SafetyAnalyst should be readily avail- temwide view of crash density, motorist risk, road per- able within highway agencies, but some effort may be formance, and potential for improvement. required to complete data assembly. SafetyAnalyst Star ratings do not require crash data and are includes a data management tool to help import and based solely on road and traffic characteristics. Star manage the necessary data inputs. ratings require as input approximately 40 key data Information provided by the participating states elements related to safety. A unique aspect of the pro- indicates that only a few have decided to incorporate tocol is that it does not require detailed, site-specific the use of SafetyAnalyst into their safety planning crash data, but relies on an inspection of roadway efforts and fewer yet plan to make the software an design features that can be done from a videolog. integral part of their efforts. In general, the comments Countermeasures selection software is also pro- provided by the states suggest that the very limited vided by usRAP and can be calibrated for applica- use of the software is due to the large data require- tion to the road network of any highway agency. ments. Minnesota staff indicated that even though the The methodology requires assembling required SafetyAnalyst data requirements were based on their data inputs (roadway and traffic characteristics) database, it took them more than a month to load and from new or existing video records while some get the model running. Minnesota staff also indicated elements may be obtained from existing roadway that they intend to use SafetyAnalyst for improving inventories. An evaluation of each location on the the identification of black spots but that the software network is then provided by the usRAP software. was not capable of assisting them with identification Crash countermeasures are identified, crash reduc- of candidates for systematic improvements. Missouri tion benefits are computed, and a benefit-cost ratio indicated that it intends to incorporate SafetyAnalyst is calculated to help prioritize the countermeasures. into its statewide safety planning efforts, is in the Nearly 70 common crash countermeasures are con- process of purchasing the license, and is working on sidered by the software, including roadway improve- making its intersection and segment characteristics ments, median treatments, shoulder paving and databases compatible with the software requirements. widening, roadside improvements, and pedestrian Iowa and North Carolina indicated that they do not and bicycle facilities. intend to use SafetyAnalyst to support their safety The software analysis tool provides a list of planning efforts and instead will continue to use and potential safety improvement projects, suggested 20

OCR for page 19
countermeasures, project location, estimated project 85th% tangent speed = 60 mph cost, estimated project benefits (in terms of fatal and serious-injury crashes reduced and in monetary terms) and benefit-cost ratio. The usRAP risk-mapping protocol has been pilot tested in eight states, and the star rating protocol has Fatal + Injury + PDO been tested and validated against crash data in two Fatal + Injury states. These pilot studies have demonstrated the technical feasibility of the usRAP risk-mapping and Bonneson et al. (2007) performance-tracking protocols for states with good Fitzpatrick et al. (2000) quality crash data. A third pilot study evaluating the application of the software and analytical processes in a state with more challenging data issues is nearing completion. Given the very limited testing to date, it Figure 11 Curve crash rates as a function of radius.9 is too early to forecast how widely this new method- ology will be deployed after the initial pilot tests. Texas State Initiatives to Develop The Texas Transportation Institute at Texas A&M New Methodologies and Tools University studied horizontal curves along Texas's A review of the safety literature combined with farm-to-market road system.8 These curves were conversations with state DOT staff and university selected based on identification as at-risk locations that researchers revealed a number of initiatives that are don't regularly show up using traditional "hot-spot" intended to fill the gap in the analytical process asso- techniques. In support of a system approach for find- ciated with identifying candidates for safety invest- ing and prioritizing the most at-risk curves, a relation- ments in rural areas. New methodologies and tools ship was developed between crash rate and curve are being developed, including statistical models and radius (see Figure 11). describing surrogates to crashes to assist with the efforts to find and prioritize at-risk locations on the Minnesota rural systems where more than one-half of severe Mn/DOT has recently published research that crashes occur, but where crash densities are very low. analyzed three components of the state's rural high- Examples of these initiatives in Iowa, Texas, and way system--horizontal curves, STOP-controlled Minnesota are discussed below. rural intersections, and two-lane highway segments. These features were selected for analysis because Iowa the data-driven process associated with Minnesota's Iowa State University is in the process of con- ducting a safety analysis of low-volume rural roads. 8 The primary objective of the project is to develop a J. Bonneson, M. Pratt, J. Miles, and P. Carlson. Development safety performance function for low-volume rural of Guidelines for Establishing Effective Curve Advisory Speeds county highways and a new statistical model. The (FHWA/TX-07/0-5439-1). Texas Transportation Institute, Texas Department of Transportation, FHWA, U.S. DOT, Octo- new model would then be incorporated into the Iowa ber 2007. Traffic Safety Data Service, which provides techni- 9 J. Bonneson, D. Lord, K. Zimmerman, K. Fitzpatrick, and M. cal assistance to county highway agencies, including Pratt. Development of Tools for Evaluating the Safety Implica- preparation of maps and lists of at-risk locations and tions of Highway Design (FHWA/TX-07/0-4703-4). Texas recommendations of potential safety improvement Transportation Institute, Texas Department of Transportation, projects.7 FHWA, U.S. DOT, 2007. K. Fitzpatrick, L. Elefteriadou, D. W. Harwood, J. M. Collins, J. McFadden, I. B. Anderson, R. A. Krammes, N. Irizarry, K. D. Parma, K. M. Bauer, and K. Pas- 7R. Souleyrette. "Safety Analysis of Low Volume Rural Roads setti. Speed Prediction for Two-lane Rural Highways, (FHWA- in Iowa" (research project in progress). Iowa State University, RD-99-171). Texas Transportation Institute, FHWA, U.S. Iowa Department of Transportation. DOT, 2000. 21

OCR for page 19
SHSP identified rural curves, intersections, and preparation of a Countywide Safety Plan for Olmsted segments as priorities based on the distribution of County, Minnesota. One of the key results of a severe crashes. The research resulted in the iden- data-driven analysis process was the finding that tification of analytical processes for identifying road-departure crashes on horizontal curves were and prioritizing at-risk locations that would be overrepresented--40% of severe road-departure candidates for the proactive deployment of low-cost crashes occurred on curves, even though curves safety improvements. made up only 15% of rural county highway mileage. The research projects identified the characteristics The methodology was used to evaluate all 241 curves of the locations with crashes and then developed a on Olmsted County's 324 miles of two-lane rural process for prioritizing these types of locations across highways. The objective of the analysis was to iden- almost 53,000 miles of the rural state and local high- tify a subset of curves that are most at risk and then to way systems based on the number of similarities with develop a low-cost safety project involving a sys- the features associated with locations with crashes. temwide deployment. Curves were ranked based on An example of this work, dealing with horizontal two primary factors and three secondary factors. The curves found the following: primary factors were radius (it was determined that There are literally thousands of curves scattered curves with radii between 500 and 1,500 ft had the across the state and county highway systems-- highest fraction of severe road-departure crashes) and it's estimated that there are over 3,000 curves serious crashes. The three secondary factors were along the state's 8,000 miles of two-lane traffic volume (volumes between 500 and 2,500 vehi- rural highways and over 26,000 curves along cles per day had the highest fraction of curve-related the 45,000 miles of rural county highways. crashes), presence of an intersection, and visual trap Curves average about 0.1 crashes per year, and (see Figure 13). The exercise resulted in the ranking slightly more than one-half of the curves have of 23 high-priority curves along the County's rural had no crashes during a 5-year study period. highway system--about 10% of all rural curves in Approximately 40% of the road-departure the County. Olmsted County subsequently used the crashes occur in curves even though curves results from this exercise to secure funding from Min- make up only about 10% of the system mileage. nesota's HSIP to proactively add chevrons at the 23 All curves are not equally at risk. high-priority curves (see Figure 14). The research dealing with STOP-controlled Consistent with the work completed by the Texas intersections and two-lane highway segments came Transportation Institute (FHWA/TX-07/0-5439-1),10 to a similar conclusion--all of these locations Mn/DOT found that radius could be used to find and along rural systems are not equally at risk. In addi- prioritize at-risk curves. The crash rate in curves with tion, a methodology based on a combination of radii greater than 2,000 ft approximates the average design features and traffic volume can be effec- rate on two-lane rural roads, but as curve radius tively used to develop a prioritized list of at-risk decreases, the crash rate increases. The crash rate at a locations that can then become candidates for safety radius of 1,500 ft is three times the system average, investment. the crash rate is four times the system average at a Scott County, Minnesota, also prepared a High- radius of 1,000 ft, and the crash rate is eleven times way Safety Plan and identified crashes at rural the system average at 500 ft (see Figure 12). This STOP-controlled intersections as one of its safety research also found that 90% of fatal crashes and 75% emphasis areas. The crashes at these intersections of injury crashes occurred on curves with radii less account for approximately 16% of all severe crashes than 1,500 ft. in the County. The challenge involved identifying A methodology based on this curve radiuscrash the most at-risk intersections--six severe crashes rate relationship was applied and refined as part of the occur annually across almost 100 rural intersections. To help identify candidates for safety improvement, 10 J. Bonneson, M. Pratt, J. Miles, and P. Carlson. Development the County conducted a prioritization exercise that of Guidelines for Establishing Effective Curve Advisory Speeds considered intersection characteristics that were (FHWA/TX-07/0-5439-1). Texas Transportation Institute, demonstrated to be associated with intersections Texas Department of Transportation, FHWA, U.S. DOT, Octo- with crashes--skewed approaches, proximity to a ber 2007. horizontal curve, traffic volume, distance from the 22