Development of Roundabout Crash Prediction Models and Methods (2019)

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.

2 Background This chapter provides background information regarding the research problem statement and objectives that framed the research need leading to these findings. It also briefly sum- marizes the tasks and research approach used in conducting this work. 1.1 Overview NCHRP Project 17-70 developed safety performance fac- tors (SPFs) and crash modification factors (CMFs) that can be used to estimate the severity and number of crashes likely to occur at roundabouts (a roundabout is defined in NCHRP Report 672: Roundabouts: An Informational Guide, 2nd edition) under various rural and urban contexts for single-lane and multilane roundabouts. This report summarizes the work conducted and the find- ings from the research. It consists of the following chapters: • Chapter 1, Background—provides the research problem statement for NCHRP 17-70 and summarizes the Phase 1 and Phase 2 work scope. • Chapter 2, Literature Review Approach and Findings— describes the scope of the literature review and presents the key findings. • Chapter 3, Framework for Safety Prediction and Data Needs—presents the framework used to identify the candidate crash prediction models and data needs to develop them. • Chapter 4, Data Collection Approach and Findings— describes the approach used to collect the crash, geometric, and traffic volume data used to develop the crash predic- tion models. It also presents characteristics of the final database. • Chapter 5, Crash Prediction Model Development Approach—describes the approach used to develop the crash prediction models and conduct the analysis within the project. • Chapter 6, Research Findings—presents the findings and final crash prediction models from the research. • Chapter 7, Conclusion and Suggested Research— summarizes the conclusion from the research and provides suggestion for further research as it relates to roundabouts. In addition, appendices provide additional background and technical information regarding the research approach. These appendices can be found by searching the TRB website for “NCHRP Research Report 888”. 1.2 Research Problem Statement The AASHTO Highway Safety Manual (HSM), 1st edition, includes SPFs—predictive models—for traditional inter- section forms (e.g., four-legged, signalized intersections) found on rural two-lane roads, rural multilane highways, and urban/suburban arterials. These prediction models allow transportation professionals to assess the benefits of pro- viding turn lanes, modifying intersection phasing, adding intersection lighting, and other such design decisions. The HSM, 1st edition, does not include crash prediction models for roundabouts; therefore, practitioners are not able to quan titatively assess the crash-reduction benefits of providing a roundabout at a specific intersection or to investigate the safety effects of design decisions at single-lane and multilane roundabouts. Research consistently indicates that roundabouts generally provide substantial reductions in crashes—in particular, reduc- tions in serious injuries and fatalities—but critical attributes that make some roundabouts more successful than others at reducing crashes are not understood. Although CMFs for roundabouts have been developed, these CMFs do not reflect the safety effects of site-specific roundabout design features. The predictive methods in the HSM Chapter 10 (Predictive Method for Rural Two-Lane, Two-Way Roads) and Chap- ter 11 (Predictive Method for Rural Multilane Highways) do C H A P T E R 1

3 not address roundabouts; predictive methods are needed for roundabouts on these roadway types. Developing crash prediction models that enable practi- tioners to understand and compare key design tradeoffs will result in better informed decisions and higher returns on safety investments. Potential users of this research include state and local agencies responsible for safety management and for planning and implementing more effective safety programs and projects. 1.3 Objectives The objective of this research was to develop SPFs and CMFs for all road users (including nonmotorized users) that can be used to estimate the severity and number of crashes likely to occur at roundabouts (a roundabout is defined in NCHRP Report 672: Roundabouts: An Informational Guide, 2nd edition) under various rural and urban contexts for single- lane and multilane roundabouts. The results of this research should enable transportation agencies to quantify the safety effects of roundabout design features and the safety perfor- mance of implementing a roundabout at a particular location so that safety performance can be compared to alternative intersection configurations for the same location. The research should answer the following three questions at a minimum: 1. How do geometric features—and combinations of features—influence the number, type, and severity of crashes at the roundabout? 2. How do operational features—and combinations of features—influence the number, type, and severity of crashes at the roundabout? 3. How do driver learning curves influence the number and severity of crashes at any age roundabout? The results of this research are presented so they can be incorporated into (1) a future edition of the HSM, (2) the AASHTO SafetyAnalyst software, and (3) the Interactive Highway Safety Design Model software. 1.4 Summary of Research Approach The research objectives for Phase 1 and Phase 2 were addressed through a work program involving a number of tasks, described in the following two sections. 1.4.1 Phase 1 Task 1: Conduct Literature Review Task 1 included (1) reviewing the literature related to roundabout safety and crash prediction, (2) identifying candidate roundabouts, (3) identifying key design consid- erations and challenges practitioners face in implementing roundabouts, and (4) establishing a clear and common ver- nacular for roundabout-related crash terminology. This task was accomplished through the following activities: • Conducting a critical review of literature, • Contacting local agencies with roundabouts, • Identifying candidate roundabout configurations, • Assessing alternatives and recommending preferred statis- tical modeling approaches, and • Reviewing roundabout-related crash definitions and devel- oping a proposed definition. This task produced working papers that described its activities and results. Task 2: Develop Data Collection Plan Task 2 was to develop a comprehensive data collection plan that consisted of (1) a framework for roundabout safety prediction and (2) a plan for collecting the data needed to develop crash prediction models for roundabouts. Task 3: Implement Data Collection Plan Task 3 was to execute the data collection plan approved in Task 2 by developing data reduction procedures, collect- ing and reducing the data, and providing status reports to the panel. Task 4: Identify Appropriate Modeling Procedure and Approach Task 4 was to prepare and implement a model develop- ment plan for the roundabout crash prediction models. This task included creating a model development plan that specified the statistical modeling approaches to be used. The candidate model development approaches were qualitatively evaluated to focus on the most-promising procedures. The model approaches were implemented to develop the initial SPFs and CMFs (or CM functions) for roundabout crash pre- diction. Task 4 also included developing an initial calibration procedure for the models and, to the extent feasible, evaluat- ing driver learning curve. Task 5: Preliminary Annotated Outline for HSM Text Task 5 was to prepare a preliminary outline of the pro- posed text to be added to the HSM.

4 Task 6: Produce Phase 2 Work Plan Task 6 was to produce and provide a Phase 2 work plan. 1.4.2 Phase 2 Task 7: Test, Validate, and Modify Modeling Procedure and Approach Task 7 was to test and refine the crash prediction models developed in Task 4. This included developing a spreadsheet tool to implement the models, testing and refining, conduct- ing a one-day practitioner workshop to further test the tool, and final refining of the models. Task 8: Annotated Outline of Final Deliverables Task 8 was to prepare an annotated outline of the final deliverables. Task 9: Proposed Text for HSM Task 9 was to propose text for the HSM that documented the roundabout crash prediction method developed for this project. This text described when, where, and how to apply the roundabout crash prediction models. Task 10: Prepare Final Deliverables Task 10 was to prepare the final deliverables for this project. Task 11: Present Research Results to AASHTO Subcommittee Task 11 was to prepare a presentation highlighting the approach taken within this project and focused on the resulting roundabout crash prediction models. Draft and final versions of this final report were prepared to document the research.