National Academies Press: OpenBook

Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis (2005)

Chapter: 3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization

« Previous: 2. Existing and Potential EDR Data Elements
Page 57
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 57
Page 58
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 58
Page 59
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 59
Page 60
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 60
Page 61
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 61
Page 62
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 62
Page 63
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 63
Page 64
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 64
Page 65
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 65
Page 66
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 66
Page 67
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 67
Page 68
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 68
Page 69
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 69
Page 70
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 70
Page 71
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 71
Page 72
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 72
Page 73
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 73
Page 74
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 74
Page 75
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 75
Page 76
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 76
Page 77
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 77
Page 78
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 78
Page 79
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 79
Page 80
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 80
Page 81
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 81
Page 82
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 82
Page 83
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 83
Page 84
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 84
Page 85
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 85
Page 86
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 86
Page 87
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 87
Page 88
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 88
Page 89
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 89
Page 90
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 90
Page 91
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 91
Page 92
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 92
Page 93
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 93
Page 94
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 94
Page 95
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 95
Page 96
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 96
Page 97
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 97
Page 98
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 98
Page 99
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 99
Page 100
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 100
Page 101
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 101
Page 102
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 102
Page 103
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 103
Page 104
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 104
Page 105
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 105
Page 106
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 106
Page 107
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 107
Page 108
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 108
Page 109
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 109
Page 110
Suggested Citation:"3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization." National Academies of Sciences, Engineering, and Medicine. 2005. Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis. Washington, DC: The National Academies Press. doi: 10.17226/21974.
×
Page 110

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.

43 3. EDR Data Needs for Roadside Safety Analyses: Identification and Prioritization 3.1 Objective The success of roadside and vehicle safety research is critically dependent upon the validity and consistency of collected crash data. To date, a majority of the accident database elements are collected or derived based on post-crash investigation. The analysis of highway crashes has often been hindered by errors in the accuracy of the collected data as well as the unmet need for data which could not be collected with traditional methods. The objective of this component of the study was to catalog and prioritize EDR data needs which support vehicle and roadside safety research and design. The specific objectives were to determine the potential of EDR technology (1) to augment data collection for existing roadside and vehicle accident databases, and (2) to support future roadside safety research needs by providing a new source of crash data previously not feasible to collect. 3.2 Methodology The following section summarizes the overall methodology followed to achieve this objective. A more detailed description of each step is provided later in this report. 1. Identify Roadside Safety Data Needs. This study pursued several avenues to methodically identify additional data elements that could be captured using EDR technology. Candidate data elements fell into two categories: (1) data elements, currently being collected manually, which could be collected by EDRs, and (2) data elements which have not been collected because the data collection capabilities of EDRs were not previously available. Analysis of Existing Accident Databases. One important use of EDR data will be replace or improve data collection for the accident databases. The research team methodically examined eight existing crash databases and three recommended database formats for candidate EDR data element needs. The databases included U.S. national accident databases, state accident databases, specialized roadside safety databases, and specialized commercial truck accident databases. The research team also examined recommended database formats or extensions including the Minimum Model Uniform Crash Criteria (MMUCC), NCHRP 350 data requirements, and NCHRP 22-15 recommended data extensions to NASS/CDS.

44 Literature Review of Roadside Safety Data Needs. The research team conducted an extensive review of the roadside safety technical literature to identify recommended improvements to data elements presently collected, and to identify data elements not presently captured that could be of significant value to the roadside safety community. 2. Develop a Catalog of Potential EDR Data Elements. Not all data elements needed for roadside safety analysis can be captured in an EDR. Fundamentally, an EDR is a vehicle-mounted device and can record only what can be measured from the vehicle. The performance of roadside features however can sometimes be inferred from the performance of the vehicle. After analysis of the data elements in each database and the technical literature, a comparison was made with the listing of current and potential EDR capabilities to ascertain potential data elements. The extraction process resulted in a catalog of elements representing the intersection of feasible EDR data elements and matching data element needs. The data elements from each of these data sources were merged into a data catalog of recommended EDR Data Elements for highway crash data analysis. 3. Prioritize Candidate Data Elements that could be collected from EDRs. Because there may be insufficient memory in an EDR to store all data elements of interest, the candidate data elements were prioritized by their importance to roadside safety analyses. This chapter presents the results of a priority ranking exercise conducted in collaboration with the AASHTO Task Force for Roadside Safety. 3.3 Literature Review of Roadside Safety Data Needs One of the crucial long-term benefits of EDRs will be their influence on highway crash safety research. The ready availability of real-world crash pulses in an EDR database will enable vehicle and roadside safety researchers to address a number of elusive research questions. Using EDR data it may be possible to conduct research to address several long-standing, and often technically controversial, issues. Many of these issues are the subject of current or previous NCHRP projects. Potential research questions on which EDR data may provide unique insights include: • How relevant are the impact conditions used in NCHRP 350? • For roadside crashes, is there a linkage between vehicle acceleration and occupant injury? How realistic is the flail space model when evaluated against actual EDR crash pulses and hospital injury records? • Are current vehicle designs compatible with current roadside safety hardware designs? • Do impacts with soft roadside safety devices, e.g. crash cushions, lead to late airbag deployments?

45 • Are advanced occupant restraint systems, e.g., dual stage inflator systems, performing as designed? • How accurate are the delta-V estimates in U.S. national accident databases? • What is the distribution of impact speeds as a function of roadside object struck? • Coupling EDR pre-impact data with highway design data, what are the relationships between highway geometric design and the probability of a runoff road event? The roadside safety literature was methodically reviewed to search for candidate EDR elements. Several previous research studies, described here, have explicitly recommended the collection of additional accident data elements to support improved roadside safety research. Although the original authors of these studies may have intended for these data to be collected with conventional accident investigation techniques, EDRs offer a promising new method of accident data collection. The availability of EDRs may allow access to data elements currently not collected, and may provide more accurate measurement of elements already being collected. The technical literature is also an excellent source of future research needs from which future data needs can be inferred. In fact, most of the current literature focuses on research needs in general, as opposed to specific data needs. Note that our goal in identifying research needs was solely to deduce additional data needs. The list of research needs presented in this chapter is by no means exhaustive. Future research efforts, e.g. the vehicle rollover problem, will undoubtedly expand on this list, and lead to new data element requirements. Table 3-1 presents a summary of the roadside safety data needs identified from the literature review. Each column indicates the source where a particular data element was suggested explicitly (designated with an “X”) or implied from research needs (designated with an “I”). An annotated bibliography of the sources used for this study is provided in the appendices. The data element needs were compared with the table of existing and potential EDR data elements. Because an EDR is vehicle-mounted, the device is, of course, limited to what can be measured on the vehicle. Table 3-2 shows that many of the recommended data elements can be obtained either from existing EDR devices or may be recorded in future EDR designs. One critical data need, which can be provided by future EDRs, was knowledge of the pre- and post-crash vehicle trajectory. Another critical research need, which can be provided by future EDRs, is the orientation of the vehicle (yaw, pitch, roll) at the time of impact. Many of the data element needs can be obtained if the EDR contains VIN. The VIN contains complete information on the vehicle make, model, year, and curb weight. When these identifiers are combined with a database such as the NHTSA Vehicle Parameter database [McCullough et al, 1995], the data needs for vehicle geometry can also be obtained. The first research which uses EDR data to study highway crash safety is now beginning to be published. The pioneering research in using EDR to study highway crash injuries

46 was performed by Kullgren et al (1995, 1998, 2000) using the Crash Pulse Recorder, a retrofit EDR developed specifically for these research studies. Gabler et al (2004) used production vehicle EDR data to validate the accuracy of delta-V estimates of crash severity. Gabauer and Gabler (2004; 2005) utilized production EDR data to evaluate the the flail space model and the acceleration severity index used as injury criteria in roadside safety hardware crash testing.

47 Table 3-1. Data Needs for Roadside Safety Analysis as expressed in the Research Literature Data Element V i n e r , J a n u a r y 1 9 9 5 M a k , 1 9 9 5 H u n t e r a n d C o u n c i l , 1 9 9 6 M a k a n d S i c k i n g , 1 9 9 4 R o s s e t a l . , 1 9 8 8 V i n e r , J u l y 1 9 9 5 R a y e t a l . , 1 9 9 5 M a k e t a l . , 1 9 8 6 H a l l e t a l . , 1 9 9 4 R a y e t a l . , 1 9 8 6 M a k , 1 9 8 3 C o u n c i l e t a l . , 1 9 9 3 M i c h i e , 1 9 9 6 P o w e r s e t a l . , 1 9 9 5 M a k e t a l , 2 0 0 0 E s k a n d a r i a n e t a l . , 2 0 0 2 P o w e r s , 1 9 9 6 R a y e t a l . , 1 9 9 8 G l e n n o n & W i l t o n , 1 9 7 4 Average Daily Traffic X X X I X Roadway Horizontal Curvature X X X I X X Roadway Vertical Alignment X X X I X X Speed Limit I Number of Lanes X X Lane Width X I X X Presence of Median X I X Median Width X I X Presence of Paved Shoulder X X X Shoulder Width X I I X Presence of Intersection X Clear Zone Width X I X Roadside Slope X X I X I I X Feature Type X X X I X X Feature Design (Dimensions) X X I X X Feature Lateral Offset X I I X X Feature Damage X X X X X X Feature Performance Assessment X I I X X X Feature Placed Properly I

48 Data Element V i n e r , J a n u a r y 1 9 9 5 M a k , 1 9 9 5 H u n t e r a n d C o u n c i l , 1 9 9 6 M a k a n d S i c k i n g , 1 9 9 4 R o s s e t a l . , 1 9 8 8 V i n e r , J u l y 1 9 9 5 R a y e t a l . , 1 9 9 5 M a k e t a l . , 1 9 8 6 H a l l e t a l . , 1 9 9 4 R a y e t a l . , 1 9 8 6 M a k , 1 9 8 3 C o u n c i l e t a l . , 1 9 9 3 M i c h i e , 1 9 9 6 P o w e r s e t a l . , 1 9 9 5 M a k e t a l , 2 0 0 0 E s k a n d a r i a n e t a l . , 2 0 0 2 P o w e r s , 1 9 9 6 R a y e t a l . , 1 9 9 8 G l e n n o n & W i l t o n , 1 9 7 4 Feature Failure Mechanism I X Vehicle Year X X Vehicle Make X X X Vehicle Model X X X Vehicle Dimensions X X X Vehicle Mass X X X Vehicle Impact Angle X X X I I X X X X X X Vehicle Impact Velocity X X I I X X X X X Vehicle Lateral Delta-V X Vehicle Longitudinal Delta-V X Vehicle Separation Angle X X Vehicle Separation Velocity X Vehicle Encroachment Angle X X X I X I Vehicle Encroachment Velocity X I I Vehicle Impact Orientation (Yaw) X I I X X Vehicle Maneuver Prior to Encroachment X X Vehicle Trajectory (after encroachment but prior to impact) X X X X Vehicle Post-Impact Trajectory I X X X X Vehicle Stability X

49 Data Element V i n e r , J a n u a r y 1 9 9 5 M a k , 1 9 9 5 H u n t e r a n d C o u n c i l , 1 9 9 6 M a k a n d S i c k i n g , 1 9 9 4 R o s s e t a l . , 1 9 8 8 V i n e r , J u l y 1 9 9 5 R a y e t a l . , 1 9 9 5 M a k e t a l . , 1 9 8 6 H a l l e t a l . , 1 9 9 4 R a y e t a l . , 1 9 8 6 M a k , 1 9 8 3 C o u n c i l e t a l . , 1 9 9 3 M i c h i e , 1 9 9 6 P o w e r s e t a l . , 1 9 9 5 M a k e t a l , 2 0 0 0 E s k a n d a r i a n e t a l . , 2 0 0 2 P o w e r s , 1 9 9 6 R a y e t a l . , 1 9 9 8 G l e n n o n & W i l t o n , 1 9 7 4 Event Description (sequence) X X X X Injury Severity X X X Accident Location Relative to Horizontal Curve (inside or outside) X Object Struck X X X X I Vehicle Damage Dimensions X X I X Vehicle Damage Location X I Lateral Extent of Encroachment X I Impact Lateral Distance (from roadway edge) X Curb Type I X Curb Height X Curb Face Slope I Vehicle Center of Gravity Location I X Pole Trajectory (subsequent to impact) I Vehicle Yaw Rate I I Roadside Soil Condition I X

50 Data Element V i n e r , J a n u a r y 1 9 9 5 M a k , 1 9 9 5 H u n t e r a n d C o u n c i l , 1 9 9 6 M a k a n d S i c k i n g , 1 9 9 4 R o s s e t a l . , 1 9 8 8 V i n e r , J u l y 1 9 9 5 R a y e t a l . , 1 9 9 5 M a k e t a l . , 1 9 8 6 H a l l e t a l . , 1 9 9 4 R a y e t a l . , 1 9 8 6 M a k , 1 9 8 3 C o u n c i l e t a l . , 1 9 9 3 M i c h i e , 1 9 9 6 P o w e r s e t a l . , 1 9 9 5 M a k e t a l , 2 0 0 0 E s k a n d a r i a n e t a l . , 2 0 0 2 P o w e r s , 1 9 9 6 R a y e t a l . , 1 9 9 8 G l e n n o n & W i l t o n , 1 9 7 4 Roadside Soil Cover X Ditch Configuration X Tire Plow Indication X Vehicle Intrusion Depth I Vehicle Intrusion Location I Vehicle Rotations (Rollover) X Vehicle Frame Rail Spread X Vehicle Frame Rail Height X Vehicle Frontal Overhang X Vehicle Bumper Height I X Key X = Explicit research need I = Implied from research needs

51 Table 3-2. Research Data Needs vs. EDR Data Element Availability Data Element Current EDR Technology Future EDR Technology Notes Vehicle Year x 1 Vehicle Make x 1 Vehicle Model x 1 Vehicle Dimensions x 1 Vehicle Mass x 1 Vehicle Impact Angle x 2,3,4,6 Vehicle Impact Velocity x Vehicle Lateral Delta-V x Vehicle Longitudinal Delta-V x Vehicle Separation Angle x 2,3,4,6 Vehicle Separation Velocity x 3 Vehicle Impact Orientation (Yaw) x Vehicle Maneuver Prior to Encroachment x Vehicle Trajectory (after encroachment but prior to impact) x 3,4,6 Vehicle Post-Impact Trajectory x 3,4,6 Vehicle Stability x 5 Vehicle Yaw Rate x Vehicle Rotations (Rollover) x Vehicle Frame Rail Spread x 1 Vehicle Frame Rail Height x 1 Vehicle Frontal Overhang x 1 Vehicle Bumper Height x 1 Notes: 1. Can be derived from VIN and associated Vehicle Parameters Database 2. Impact angle and separation angle can be reconstructed by combining the vehicle trajectory with site measurement of the barrier location and orientation. 3. Will require longer recording time to capture the entire event including separation 4. Requires extension of current practice of recording pre-crash parameters to encompass the post-crash period 5. Assumes stability can be inferred from yaw, pitch, roll vs. time 6. Determination of the vehicle trajectory requires measurement of (a) acceleration along the longitudinal, lateral, and vertical axes (b) vehicle yaw, pitch, and roll rates, and (c) final resting position/orientation of the vehicle (obtained from site inspection). For planar collisions without rollover, trajectory can be obtained from a more limited set of parameters: (a) acceleration along the longitudinal and lateral axes (b) the vehicle yaw rate, and (c) final resting position/orientation of the vehicle (obtained from site inspection).

52 3.4 Examination of Existing Accident Databases One of the most important near-term uses of EDR data will be to improve the collection of data for existing accident databases. This section describes the methodology, analysis, and results of a study to determine the potential of using EDR data to augment data collection for roadside and vehicle crashes. Our analysis examined the following wide spectrum of major accident databases, crash test databases, and recommended database formats: • Fatality Analysis Reporting System (FARS) • National Automotive Sampling System/Crashworthiness Data System (NASS/CDS) • National Automotive Sampling System/General Estimates System (NASS/GES) • Highway Safety Information System (HSIS) • Longitudinal Barrier Special Studies (LBSS) • Model Minimum Uniform Crash Criteria (MMUCC) • NHTSA Vehicle Crash Test Database • NCHRP Report 350 Roadside Feature Performance Test Elements • NCHRP 22-15 Recommended NASS/CDS Data Elements • Trucks Involved in Fatal Accidents (TIFA) • Motor Carrier Management Information System (MCMIS) – Crash File Our objective was to consider a broad array of database types to determine the potentially large range of EDR data uses. FARS, NASS/GES, and NASS/CDS are extensive U.S. accident databases. HSIS and LBSS are specialized highway and roadside safety databases. The NCHRP 22-15 data elements are recommended extensions to NASS/CDS to better capture the performance of roadside features. TIFA and the MCMIS crash file are specialized heavy truck and bus databases. MMUCC is a recommended database format to coordinate data collection efforts across states and thus is critical due to the potentially wide application of EDR technology. The NHTSA Vehicle Crash Test Database Protocol and the NCHRP Report 350 Roadside Feature Performance Test Protocol are detailed descriptions of data for vehicle and roadside hardware testing, respectively. Barring limitations on instrumentation and data collection, these protocols are assumed to represent an ideal set of information. The NHTSA Crash Test Database for example contains the complete description and results of over 5000 vehicle crash tests conducted since the late 1970s. Because an EDR is in many ways analogous to the instrumentation used in laboratory crash tests, the NHTSA Crash Test Database and the NCHRP Report 350 are invaluable guides to the data required to perfectly describe a real world accident. To better facilitate an understanding of the current state of accident data collection and how EDR technology may potentially augment current data collection, a data element classification methodology was developed. Due to its wide application in crash analysis, the FARS database was chosen as the test database for the methodology development.

53 All examples in the methodology refer to elements in the FARS database; this methodology has been applied to the remaining databases and collection protocols. 3.4.1 Classification Methodology Due to the number of elements and range of data present in the existing databases, the research team recognized the need for a methodical approach to the classification of the data elements within each database. The classification scheme must provide a more definitive means of identification of candidate data elements as well as provide a measure of simplification for discussion and presentation purposes. A modified Haddon approach was chosen and is shown in Table 3-3. Table 3-3. Modified Haddon Matrix Pre-Crash Crash Post-Crash Time-Invariant Human Vehicle Environment In this approach, the crash event has three components (1) the Human, (2) the Vehicle, and (3) the Environment. The Human is typically the crash victim and includes both vehicle occupants as well as non-occupants such as pedestrians. The Environment includes entities outside the human-vehicle pairing which may have influenced either the actions leading to the crash or the outcome of the crash. Examples of Environment factors would be the weather, road curvature, or guardrail systems. The event can be further broken down into three time phases: pre-crash, crash, and post- crash. Each event phase has a characteristic time duration. The characteristic duration of the pre-crash phase is generally measured in seconds to minutes. The duration of the crash duration is measured in milliseconds to seconds. The characteristic duration of the post- crash phase is measured in minutes to hours. The objective in conducting this classification exercise is to identify data elements that could potentially be collected more accurately or more effectively by systems such as an EDR. Our approach is to classify each data element both by where and when each data element could be measured by an electronic system such as an EDR. The location where a data element could be measured will include the occupant, the vehicle, or the environment. The time when a data element could be measured is the earliest time that a data element could be captured. Unlike an EDR that can capture some data elements in the pre-crash phase, current accident investigation must collect all data elements in the post-crash phase. Note that this classification methodology is merely used as an analytical tool for seeking links between accident databases and EDR elements. Other methods may be more appropriate for other research objectives. For example, as a given crash may involve

54 several vehicles and a number of persons, accident records are generally organized as relational databases of linked tables. FARS for example has three tables – (1) accident, (2) vehicle, and (3) person – with appropriate linking identifiers. A complete description of the crash using the Haddon Matrix approach would require a multi-layer Haddon Matrix with a separate matrix for each person-vehicle combination. Classification Guidelines Definitions for the corresponding rows and columns in the modified Haddon Matrix are illustrated below. Again, all variables used as examples are elements from the FARS database. • Time-Based Category Definitions. Categorization by time or crash phase allows a lengthy list of data elements to be disaggregated by when the elements could be measured by a system such as an EDR. ¾ Pre-Crash Variables. Data elements pertaining to the time prior to the event (not including Time Invariant Variables). ¾ Crash Variables. Descriptors of the crash event and the surrounding environment at the time of the crash. ¾ Post-Crash Variables. Characterization variables pertaining to the time after the incident. ¾ Time Invariant Variables. Time invariant variables are fixed over the characteristic time of the event – loosely defined here to be one hour before and after the actual crash. Time-invariant variables would include a) fixed identifiers that are not time-dependant (e.g. VIN), (b) historical data (i.e. previous traffic violations), and (c) data elements unlikely to change over the course of the entire event sequence, (e.g. driver weight) • Location-Based Category Definitions ¾ Human Variables. This category includes identifiers that apply specifically to the human(s) involved in the incident. These identifiers may be a quantitative measure (i.e. driver weight) or may be based on a human judgment system (i.e. license status or regulatory compliance). Human variables would also include the response of the human(s) to the crash event (i.e. injury severity). ¾ Vehicle Variables. This category includes (a) descriptive identifiers applying specifically to the vehicle(s) involved in the incident (i.e. vehicle body type), and (b) response of the vehicle(s) to the crash event (i.e. airbag deployment) ¾ Environmental Variables. Descriptive identifiers applying specifically to the surroundings of the crash location.

55 • Additional Classification Conventions. The assumption that electronic instrumentation is available is based on the most probable method of instrumentation. For example, the AVOID variable describes the maneuver initiated by the driver to avoid the crash. Although it may be possible to instrument the driver of the vehicle in the future, a more plausible method of assessment would be to instrument the vehicle to indicate the steering and/or braking applied by the driver prior to the crash. Thus, this variable is classified as a pre-crash vehicle attribute. Examination Results and Extraction of Potential Data Elements After the data element classification was performed for a targeted database, a comparison was made with the listing of current and potential EDR capabilities to ascertain potential data elements. The extraction process resulted in a catalog of elements representing the union of feasible EDR data elements and matching elements present in each target database. The classification and potential data element extraction results for each of the target databases are presented below. A brief description of each database is provided along with the data elements. Referenced tables use the following general tabular anatomy: • A brief description of each variable in the database including information that is pertinent to its classification location. • The source database table (if applicable). • Major column divisions correspond to three “where” components (1) the Human crash victim, (2) the Vehicle, and (3) the Environment. Note that these were chosen as the major division based on the assumption that most of EDR elements will be contained within the Vehicle category. • Each major column is divided into four minor columns corresponding to the time classifications: pre-crash, crash, post-crash, and time in variant. • An indication of whether the variable is derived from other data. Note that these variables are not classified since each is generated from other information collected within the framework of the modified Haddon approach. For example, the VEH_NO variable is a unique numeric value used to identify each vehicle involved in a given accident case. As this value is simply an arbitrary vehicle identifier assigned by the database coders, it is listed as a derived variable. 3.4.2 FARS FARS is a comprehensive census of all traffic related fatalities in the United States, which has been maintained and operated by the National Highway Traffic Safety Administration (NHTSA) since 1975 [Tessmer, 1999]. For a case to be included in this

56 database, it must involve a motor vehicle traveling on a primarily public roadway and death of an individual involved within 30 days of the incident. Each incident is characterized by the collection of approximately 175 data elements split among an accident table, a vehicle table, and a person table. Extraction of Potential EDR data elements from FARS A complete listing and classification of the variables in the FARS database is provided in the appendices. The variables have been grouped according to the FARS table system, and arranged in alphabetical order within each table. Note that variables occurring in multiple tables are sorted alphabetically at the end of the listing, with a corresponding annotation describing in which tables they appear. As the FARS database was not created based on the Haddon classification methodology, there are inherent nuances in the application of our methodology to this database. One particular nuance is the elements that span more than one category in either the time or characteristic dimension. For instance, the Driver Presence Variable, DR_PRES, has the following entry possibilities: (1) Driver operated vehicle, (2) Driver left scene, (3) No driver, and (4) Unknown. This variable spans the time dimension from pre-crash to post- crash as the earliest measure of whether the driver is operating the vehicle can be obtained in the pre-crash phase and the earliest measure of whether the driver leaves the scene can be obtained in the post-crash phase. In the other direction, the Notification Variables, NOT_HOUR and NOT_MIN, indicate the time of notification for need of medical services and span both the Human and Vehicle categories. The Human component would correspond to time of a 911 call placed for the particular accident while the Vehicle component would correspond to a vehicle equipped with an ACN system that sends the notification for the need of medical services. Link between EDR and FARS To complete the linkage between the potential EDR data elements and the variables captured in the FARS database, the classification methodology has been applied to the listing of potential EDR elements. By classifying both the EDR data and FARS database elements using the Haddon matrix and then matching the two classified lists, the FARS elements presented in Table 3-4 were identified as potential candidates for EDR data based on the current and future EDR technology. Note that the extraction results presented in this report are based on the assumption that there is no widespread database available to link GPS information to roadway data such as milepost, route number, and posted speed limit. This assumption is also applicable to all subsequent databases analyzed.

57 Table 3-4. FARS-EDR Compatibility Variable Name Variable Description Current Technology Future Technology Notes TRAV_SP Estimation of vehicle travel speed X DR_CF1 Driver related factors (often indicates the cause of the crash) X X 1 DR_CF2 Driver related factors (often indicates the cause of the crash) X X 1 DR_CF3 Driver related factors (often indicates the cause of the crash) X X 1 DR_CF4 Driver related factors (often indicates the cause of the crash) X X 1 DR_PRES Driver presence (Driver operated vehicle, driver left scene, no driver, etc.) X 2 VEH_CF1 Vehicle related crash factor (often indicates crash cause) X X 3 VEH_CF2 Vehicle related crash factor (often indicates crash cause) X X 3 MAN_COL Manner of collision (head on, rear end, etc.) X AXLES Total number of axles on the vehicle X 4 MODEL Vehicle model (see MAK_MOD) X 4 VIN Vehicle identification number (up to the first 12 digits) X 4 VIN_# (1-12) xth number of the VIN X 4 VIN_LNGT Actual length of the VIN number for the vehicle X 4 BODY_TYP Indicates vehicle body type based on NHTSA classification X 4 VINA_MOD Model of vehicle as obtained by the VINA program X 4 VIN_BT Vehicle body type from VINA program X 4 VIN_WGT Weight of the vehicle (excluding trucks) X 4 WHLBS_LG Longest wheelbase for the model vehicle (based on VINA program) X 4 WHLBS_SH Shortest wheelbase for the model vehicle (based on VINA program) X 4 SER_TR Truck version of VIN_BT (obtains vehicle body type) X 4 AIR_BAG For vehicle occupants, indicates whether air-bag deployed X REST_USE Indicates the type of restraint used X 4 DAY Day of the month of the crash X DAY_WEEK Day of the week of the crash (calculated from other date/time information) X YEAR Year that the crash took place X ROLLOVER Indicates whether a rollover occurred and if it was the first event or a subsequent event X HOUR Hour when the crash occurred X MAKE Indicates the make of the vehicle X 4 MAK_MOD Make information concatenated with the model information X 4 MINUTE Minute of when the crash occurred X

58 Variable Name Variable Description Current Technology Future Technology Notes MOD_YEAR Indicates the model year of the vehicle X 4 MONTH Month when the crash occurred X EMER_USE Indicates whether the vehicle was in emergency use at the time of the crash X 10 DR_WGT Indicates the weight of the driver in pounds X 2 SCH_BUS Indicates whether the accident involved a school bus functioning as such X 5 PER_TYP Indicates situation of occupant (driver, passenger of vehicle in motion, passenger of vehicle not in motion, etc.) X 2 OCCUPANTS Actual number of occupants in the vehicle at the time of the crash X 2 WEATHER Indicates atmospheric conditions at the time of crash X 6 CITY City code based on GSA codes X 7 LATITUDE Global position of the crash location (latitude) X 7 LONGITUD Global position of the crash location (longitude) X 7 COUNTY County of incident based on GSA codes X 7 STATE State where the crash occurred (GSA codes) X 7 LGT_COND Lighting condition at the scene of the crash X 5 NOT_HOUR Hour of notification for the need of medical services X 8 NOT_MIN Minute of notification for the need of medical services X 8 SEAT_POS Indicates the seating position of a particular occupant X 2 AVOID Driver executed maneuver to attempt to avoid the crash X 9 SUR_COND Surface conditions at crash site (i.e. wet, dry, snow, etc.) X 9 Notes: 1. Some choices such as (45) Driving less than posted maximum can be inferred using current EDR technology while others like (56) Improper tire pressure may be inferred from future EDR technology. 2. Assumes a weight sensor in all vehicle seating positions. 3. Some choices such as (17) Airbag can be obtained from current EDR technology while others such as (12) Wipers may be inferred from future EDR technology. 4. Can be derived from VIN and associated Vehicle Parameters Database 5. May be inferred from usage of lights 6. May be inferred from usage of windshield wipers 7. Can be derived assuming GPS position data is recorded 8. Assuming implementation of ACN system 9. May be inferred from precrash information from ABS systems or steering or braking information 10. May be inferred from use of a siren. In merging the classified data elements, the following categories emerged with respect to how EDR data may be useful to the FARS database:

59 Direct: EDR data can be obtained directly from the EDR and transferred to the database without any intermediate inference (note that mathematical operations to convert EDR raw data are ignored). Direct elements would include vehicle travel speed (TRAV_SP), time of the crash (HOUR, DAY, MINUTE), and deployment of air bags during the event (AIR_BAG). These appear to have the largest potential for increasing the accuracy and efficiency of the information present in accident databases. Indirect: EDR data can be utilized to infer or derive a particular variable captured in a current accident database. An example would be the MILEPOINT variable that identifies the roadway mile point location of a particular crash event. Assuming that EDR has GPS capability and the roadway databases contain sufficient detail, one could determine the mile point of the crash from the EDR GPS latitude and longitude. Partial Direct: EDR data can be utilized directly to fulfill a portion of a currently captured variable. A possible example could be the Related Factors – Vehicle Level variable (VEH_CF#), which indicates malfunctions in the vehicle that could have attributed to the crash event. Assuming EDR data could directly indicate the functionality of the brake system during the crash, this information could be used to choose or eliminate (02) Brake System as a crash related factor. It is unlikely, however, that EDR data will be able to directly determine a (31) Hit-and-Run Vehicle; thus, information from the EDR can only partially fulfill the requirements of the variable. Partial Indirect: EDR data can be utilized to infer or derive a portion of a particular variable captured in a current accident database. Example variables in this category include all crash related factors. Consider the Related Factors - Driver Level variable (DR_CF#), in which the data coder has approximately one hundred possible choices for factors related to the cause of the crash. Two possible choices include (10) Deaf and (44) Driving too Fast for Conditions or in Excess of the Posted Speed Limit. An inference made from the EDR data may be useful in assessing whether the crash is speed related. It is unlikely, however, that EDR data will be useful in ascertaining whether the driver was deaf. Thus, EDR data could only be used to infer some of the subsets (choices) within a particular variable. 3.4.3 NASS/CDS NASS/CDS provides a detailed record of a national sample of approximately 5,000 crashes investigated each year by NHTSA at 27 locations throughout the United States [NHTSA, 2000]. This database includes a random sample of minor, serious and fatal crashes involving cars, light trucks, vans and sport utility vehicles. Compared with the FARS and GES database, the data collected in NASS/CDS is much more detailed and includes approximately 400 data elements. Table 3-5 presents the results of the examination of the NASS/CDS database and subsequent extraction of potential EDR data elements.

60 Table 3-5. NASS/CDS Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes DAYWEEK Day of week of the accident X EVENTS Number of recorded events in accident X MANCOLL Manner of collision X MONTH Month of accident X TIME Time of accident X YEAR Year of accident X ANGTHIS Heading angle for this vehicle X ANTILOCK Antilock brakes X BAGDEPFV Air bag deployment, first seat frontal X BAGDEPOV Air bag deployment, other X BODYTYPE Vehicle body type X 1 CARBUR Carburetion X 1 CURBWGT Vehicle curb weight X 1 DRIVE Front/rear wheel drive X 1 DRPRES Driver presence in vehicle X 2 DVEST Estimated highest delta v X DVLAT Lateral component of delta v X DVLONG Longitudinal component of delta v X DVTOTAL Total delta v X FOURWHDR Four wheel drive X 1 FRTWHLDR Front wheel drive X 1 IMPACTSP Impact speed X LGTCOND Light conditions X 3 MAKE Vehicle make X 1 MANEUVER Attempted avoidance maneuver X 6 MODEL Vehicle model X 1 MODELYR Vehicle model year X 1 PREISTAB Pre-impact stability X 5 RESTYPE Restraint type X 1 ROLINDIR Direction of initial roll X ROLLOVER Rollover X SERTR VIN series truck X 1 SURCOND Roadway surface condition X 6 VEHTYPE Type of vehicle X 1 VEHUSE Vehicle special use (This trip) X 8 VEHWGT VIN vehicle weight X 1 VIN Vehicle identification number X 1 VINAMOD VIN model cars & trucks X 1 VINBT VIN body type X 1 VINLNGTH VIN length X 1 VINMAKE VIN make X 1 VINMODYR VIN model year X 1 WEATHER Atmospheric conditions X 4 ABELTAVL Automatic belt system availability/function X ABELTUSE Automatic belt (passive) system use X ABELTYPE Automatic (passive) belt system type X BAGAVAIL Air bag system availability X BAGAVOTH Other frontal air bag availability/function X

61 Variable Name Variable Description Current Technology Future Technology Notes BAGDEPLY Air bag system deployed X BAGDEPOT Other air bag system deployment X BAGFAIL Air bag system failure X DVBAG Longitudinal component of delta v for airbag deployment X MANAVAIL Manual belt system availability X MANUSE Manual belt system use X ROLE Occupant's role X 2 SEATPOS Occupant's seat position X 2 WEIGHT Occupant's weight X 2 DOF1 Direction of force (highest) X DOF2 Direction of force (2nd highest) X PDOF1 Clock direction for principal direction of force in degrees (highest) X 7 PDOF2 Clock direction for principal direction of force in degrees (2nd highest) X 7 WHEELBAS Original wheelbase X 1 Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- May be inferred from usage of lights 4- May be inferred from usage of windshield wipers 5- Assumes stability can be inferred from yaw, pitch, roll vs. time 6- May be inferred from precrash information from ABS systems or steering or braking information 7- Assuming accelerometers in perpendicular horizontal directions 8- May be inferred from use of a siren. The reader should refer to the appendices for a complete listing and classification of the variables contained in the NASS/CDS database. Note that the table is organized in a manner consistent with the forms present in the NASS/CDS collection format. 3.4.4 NASS/GES The function of NASS/GES is to present a representative sample of all police-reported motor vehicle accidents in the United States [NHTSA, 2001]. Criteria for selection include the involvement of a motor vehicle traveling on a public road and a crash that results in property damage, occupant injury, a fatality, or any combination of these outcomes. The accident reports are sampled from approximately 400 police jurisdictions in 60 areas across the United States. Each area is selected with the intent of providing a spectrum that is indicative of geographical, roadway, population and traffic characteristics in the entire country. Complexity of collected crash information is analogous to that of the FARS database and includes approximately 130 data elements. Table 3-6 presents the results of the examination of the NASS/GES database and subsequent extraction of potential EDR data elements.

62 Table 3-6. NASS/GES Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes MONTH Month in which the crash occurred X YEAR Year in which the crash occurred (four digits) X WEEKDAY Day of the week in which the crash occurred X HOUR Hour in which the crash occurred X MINUTE Minute in which the crash occurred X SUR_COND Roadway surface condition at the time of the crash X 5 LGHT_CON General light conditions at the time of the crash X 3 WEATHER General description of atmospheric conditions at the time of the crash X 4 MAKE Indicates the make of the motor vehicle involved in the crash X 1 MODEL Indicates the model of the motor vehicle involved in the crash X 1 MODEL_YR Indicates the model year of the involved vehicle X 1 VIN First 11 characters of the Vehicle Identification Number X 1 EMCY_USE Indicates if the vehicle was in emergency use at the time of the crash X 3 NUMOCCS Indicates the number of persons (including drivers) within an involved vehicle X 2 SPEED Speed of involved vehicle prior to event (miles per hour) X FACTOR Indicates vehicle related factors that may have contributed to the crash (only one is coded) X 5 ROLLOVER Indication of a rollover for an involved vehicle (includes tripping mechanism) X P_CRASH1 Description of the vehicle's activity just prior to the crash X 5 P_CRASH3 Describes the driver actions in response to the impending crash (i.e. steering, braking, etc.) X 5 P_CRASH4 Assessment of the stability of the vehicle just after the corrective action but prior to the initial impact X 6 P_CRASH5 Identifies the path of the vehicle prior to its involvement in the crash X 5 DR_PRES Indicates the presence of the vehicle driver (used to identify driverless vehicles) X 2 SPEEDREL Indicates whether speed was a contributing factor in the crash X PER_TYPE Indicates the role of the person in the vehicle X 2 SEAT_POS Indicates the location of the occupants within the vehicle X 2 REST_SYS Indicates the occupant's use of available restraints within the vehicle X AIRBAG Indicates the presence of an airbag and it's function during the event X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions 3- May be inferred from usage of lights

63 4- May be inferred from usage of windshield wipers 5- May be inferred from precrash information from ABS systems or steering or braking information 6- Assumes stability can be inferred from yaw, pitch, roll vs. time The reader should refer to the appendices for a complete listing and classification of the variables contained in the NASS/GES database. Note that the variables have been split according to the NASS/GES table system with a separate division for variables occurring in multiple tables. 3.4.5 HSIS HSIS is a multi-state database that contains crash, roadway inventory, and traffic volume data for a select group of States [FHWA, 2000a, 2000b, 2000c, 2000d, 2001]. Created by the Federal Highway Administration (FHWA), this database fuses the information provided by each state to generate a homogeneous set of approximately 120 data elements. Table 3-7 identifies the states involved in the HSIS database and corresponding data contributions. Table 3-7. Summary of HSIS Data Available State First Year Available Average Crashes/Year Roadway Mileage Information Provided California 1991 160,000 16,000 Crash, Roadway Inventory, Traffic Volume, Intersection, Interchange/Ramp Illinois 1985 142,000 16,000 Crash, Roadway Inventory, Traffic Volume, Curve and Grade, VIN Maine 1985 39,000 23,000 Crash, Roadway Inventory, Traffic Volume, Interchange/Ramp Michigan 1985 136,000 9,700 Crash, Roadway Inventory, Traffic Volume, VIN, Guardrail/Barrier, Interchange/Ramp Minnesota 1985 77,000 134,000 Crash, Roadway Inventory, Traffic Volume, Intersection North Carolina 1991 116,000 35,000 Crash, Roadway Inventory, Traffic Volume, VIN Ohio 2002 - - (To be added in 2002) Utah 1985 46,000 14,000 Crash, Roadway Inventory, Traffic Volume, Curve and Grade, VIN Washington 1991 35,000 8,600 Crash, Roadway Inventory, Traffic Volume, Curve and Grade, Interchange/Ramp Table 3-8 presents the results of the examination of the HSIS database and subsequent extraction of potential EDR data elements.

64 Table 3-8. HSIS Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes ACC_DATE Accident date X ACCYR Accident year X DAYMTH Day of month X HOUR Hour of occurrence X MONTH Month of accident X WEEKDAY Day of week X LIGHT Light condition X 3 RDSURF Surface road condition X 5 WEATHER Weather condition X 4 CONTRIB1 Accid contrib factor(s) X 5 CONTRIB2 Accid contrib factor(s) X 5 CONTRIB3 Accid contrib factor(s) X 5 VEHTYPE Vehicle type X 1 VEHYR Vehicle year X 1 VIN VIN number X 1 MISCACT1 Driver intent X 5 SEATPOS Occupant position in vehicle X 2 REST1 Occupant safety equipment X COUNTY County X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- May be inferred from usage of lights 4- May be inferred from usage of windshield wipers 5- May be inferred from precrash information from ABS systems or steering or braking information The reader should refer to the appendices for a complete listing and classification of the variables contained in the HSIS database. Note that the variables have been split according to the data categories used in the HSIS database and that each variable is accompanied by an indication of the state(s) which provide information for that variable. 3.4.6 Longitudinal Barrier Special Study (LBSS) The LBSS was developed within NASS to examine accidents involving longitudinal barriers [Erinle, 1994]. Collected between 1982 and 1986, data exists for a total of 1,146 NASS cases. Approximately 250 data elements are collected and organized into six separate data files: the accident data file, barrier accident file, barrier contact file, driver data file, occupant data file, and the vehicle data file. Table 3-9 presents the results of the examination of the LBSS database and subsequent extraction of potential EDR data elements.

65 Table 3-9. LBSS Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes LGTCOND Light conditions at the time of the accident X 3 MANCOLL Manner of collision based on first harmful event X 5 TIME Time of the accident X WEATHER Atmospheric conditions at the time of the accident X 4 B62 Impact angle X 5 B63 Vehicle yawing angle at impact X 5 B64 Impact speed X B65 Separation angle X 5,6 B67 Post-impact trajectory X 6 B69 Rollover X B71 Confidence of impact angle X B72 Confidence of yawing angle at impact X B73 Confidence of separation angle X 6 B74 Confidence of final rest distance X 6 AVOIDMAN Attempted avoidance maneuver X 7 OCUPANTS Number of occupants (this vehicle) X 2 SURCOND Roadway surface condition X 7 AUTAVAIL Passive restraint system - availability X AUTFNCT Passive restraint system - function X MANAVAIL Active restraint system - availability X 1 MANUSE Active restraint system - use X ROLE Occupant's role X 2 SEATPOS Occupant's seat position X 2 BODYTYPE Vehicle body type X 1 CURBWGT Vehicle curb weight X 1 DOF1 Direction of force (highest) X 5 DRIVE Front/rear wheel drive X 1 DVLAT Lateral component of delta V X DVLONG Longitudinal component of delta V X DVTOTAL Total delta V X GVWR Gross vehicle weight rating X 1 MAKE Vehicle make X 1 MODEL Vehicle model X 1 MODELYR Vehicle model year X 1 ROLLOVER Rollover involvement X SPECUSE Vehicle special use (this trip) X 3 TRAVELSP Vehicle travel speed X WHEELLNG Wheelbase long X 1 WHEELSHT Wheelbase short X 1 YEAR Year of accident X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- May be inferred from usage of lights 4- May be inferred from usage of windshield wipers 5- Assumes horizontal accelerometers in perpendicular directions 6- Requires extension of recording to post-crash time period

66 7- May be inferred from precrash information from ABS systems or steering or braking information The reader should refer to the appendices for a complete listing and classification of the variables contained in the LBSS database. The variables have been split according to the file system used in the LBSS database. Note that variables occurring in multiple tables are sorted alphabetically at the end and the corresponding tables in which they appear are identified. 3.4.7 Model Minimum Uniform Crash Criteria (MMUCC) The Model Minimum Uniform Crash Criteria (MMUCC) is a minimum set of 75 crash data elements with standardized definitions that are relevant to injury control, highway and traffic safety. NHTSA and FHWA in collaboration with the Governor’s Highway State Association (formerly the National Association of Governors’ Highway Safety Representatives) to develop model minimum uniform crash criteria [NHTSA and FHWA, 1998]. MMUCC was developed in response to studies that have concluded that the use of state crash data is often hindered by the lack of uniformity between and within states. The MMUCC was developed with extensive input from the states DOT’s, and has been officially adopted by a number of states. Table 3-10 presents the results of the examination of the MMUCC protocol and subsequent extraction of potential EDR data elements. Table 3-10. MMUCC Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes C2 Crash Date and Time X C3 Crash County X C4 Crash City/Place X C8 Manner of Crash/Collision Impact X C11 Weather Condition X 4 C12 Ambient Light X 3 C13 Road Surface Condition X 5 V4 Vehicle Make X 1 V12 Gross Vehicle Weight Rating of Power Unit X 1 V13 Total Occupants In Vehicle X 2 V15 Emergency Use X 3 V19 Direction of Travel Before Crash X V21 Vehicle Maneuver/Action X 5 V25 Direction of Force to Vehicle X 6 P3 Person Type X 2 P6 Seating Position X 2 P7 Occupant Protection System Use X P8 Air Bag Deployed X P14 Contributing Circumstances, Driver X 5 CD8 Day of Week X VL1 Vehicle Identification Number X 1 VD1 Vehicle Model Year X 1

67 Variable Name Variable Description Current Technology Future Technology Notes VD2 Vehicle Model X 1 VD3 Vehicle Body Type X 1 Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- May be inferred from usage of lights 4- May be inferred from usage of windshield wipers 5- May be inferred from precrash information from ABS systems or steering or braking information 6- Can be estimated for EDRs that record acceleration on multiple axes The reader should refer to the appendices for a complete listing and classification of the variables contained in the MMUCC. The variables have been organized according to the MMUCC variable type system structure. 3.4.8 NHTSA Vehicle Crash Test Database Protocol (VEHDB) NHTSA collects detailed engineering data which describes the impact response of vehicles subjected to a crash test. The data is typically used to evaluate the ability of the vehicle structure and restraints to protect the occupants during a crash. Approximately 180 data elements are collected per test and stored in the NHTSA Vehicle Crash Test Database [ISSI, 2001]. Data elements include the acceleration profile of the vehicle and force and acceleration data for various components of anthropomorphic crash test dummies. Table 3-11 presents the results of the examination of the NHTSA Vehicle Crash Test Database (VEHDB) and the subsequent extraction of potential EDR data elements. Table 3-11. NHTSA VEHDB Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes TKCOND Description of the test track condition X TEMP Temperature at the test location at the time of the test X IMPANG Impact angle (magnitude of the angle between the longitudinal axis of vehicle 2 and the longitudinal axis of vehicle 1 or barrier in a clockwise direction) X 3 MAKE Manufacturer of the vehicle X 1 MODEL Model of the test vehicle X 1 YEAR Model year of the test vehicle X 1 BODY Test vehicle body type X 1 VIN Vehicle Identification Number as assigned by the manufacturer X 1 ENGINE Engine type of the vehicle X 1

68 Variable Name Variable Description Current Technology Future Technology Notes ENGDSP Test vehicle engine displacement (liters) X 1 WHLBAS Measured or published value for the vehicle or impactor's wheelbase X 1 VEHLEN Measured or published value for the length of the vehicle or impactor X 1 VEHWID Maximum width of the vehicle or impactor X 1 VEHSPD Resultant speed of the vehicle immediately before impact X PDOF Principal direction of force - angle between the vehicle's longitudinal axis and the impulse vector (clockwise is positive) X 3 OCCLOC Indication of the location of the test occupant in the vehicle X 2 OCCWT Weight of the non-dummy test occupant X 2 RSTTYP Type of restraint system in use at a given occupant location X DEPLOY Describes deployment performance of inflatable restraints (or firing of belt pretensioners) X SENTYP Indicates the type of sensor used for collecting measurements X SENATT Indication of where the sensor is attached (i.e. right A pillar, engine, etc.) X AXIS Axis direction for sensors measuring vector quantities X YUNITS Unit used to measure the signal of the sensor data X INIVEL Initial velocity of the sensor (linear accelerometers) X DELT Time increment of the measurement (microseconds) X DASTAT Indicates the status of the data as it appears in the submission (indicates signal validity) X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- Assumes horizontal accelerometers in perpendicular directions The reader should refer to the appendices for a complete listing and classification of the variables contained in the NHTSA Vehicle Crash Test Database. 3.4.9 NCHRP Report 350 Roadside Feature Performance Test Elements These roadside hardware performance testing procedures are the current standard utilized in the industry for longitudinal barriers, crash cushions, breakaway devices, truck- mounted attenuators and work zone traffic control devices. Analogous to the NHTSA VEHDB, NCRHP 350 is used to assess the performance of roadside safety hardware for a given set of standardized conditions. Approximately 100 data elements are collected

69 which describe the physical characteristics and dynamic performance of both the vehicle and the tested device [Ross et al, 1993]. Table 3-12 presents the results of the examination of the NCHRP Report 350 protocol and subsequent extraction of potential EDR data elements. Table 3-12. NCHRP Report 350 Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes N/A Date of the test X N/A Vehicle Identification Number X 1 N/A Vehicle make X 1 N/A Vehicle model X 1 N/A Vehicle model year X 1 N/A Mileage just prior to test X N/A Tire inflation pressure X N/A Engine Type X 1 N/A Engine Cylinder Inside Diameter X 1 N/A Curb mass - total X 1 N/A Vehicle impact speed X N/A Vehicle exit speed X 3 N/A Vehicle impact angle X 2 N/A Vehicle exit angle X 2,3 N/A Vehicle acceleration X N/A Vehicle trajectory X 3 N/A Vehicle roll rate throughout event X 3 N/A Vehicle yaw rate throughout event X 3 N/A Vehicle pitch rate throughout event X 3 N/A Final rest position of test vehicle X N/A Occupant Impact Velocity (X-direction) X N/A Occupant Impact Velocity (Y-direction) X 2 N/A Theoretical Head Impact Velocity X 2 N/A Ridedown acceleration (X-direction) X 3 N/A Ridedown acceleration (Y-direction) X 2,3 N/A Post-Impact Head Deceleration X 2,3 N/A Acceleration Severity Index X 2,3 N/A Post-impact max vehicle roll angle X 3 N/A Post-impact max vehicle pitch angle X 3 N/A Post-impact max vehicle yaw angle X 3 Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes horizontal accelerometers in perpendicular directions 3- May require extension of recording to post-crash time period The reader should refer to the appendices for a complete listing and classification of the variables contained in the NCHRP Report 350 Roadside Feature Performance Test Protocol.

70 3.4.10 NCHRP 22-15 Recommended NASS/CDS Data Elements The main objectives of NCHRP Project 22-15, entitled “Improving the Compatibility of Vehicles and Roadside Safety Hardware”, were to investigate the compatibility between vehicles and roadside safety hardware and to assess opportunities and barriers to improving compatibility. While fulfilling these objectives, Eskandarian et al (2002) suggested an improved NASS/CDS data collection form to facilitate further research in the area of vehicle-hardware compatibility. The revised data collection sheets are comprised of approximately 60 elements focused mainly on the physical and performance characteristics of the impacted roadside safety hardware. Table 3-13 presents the results of the examination of the NCHRP 22-15 Recommended NASS/CDS Data Elements and subsequent extraction of potential EDR data elements. Table 3-13. NCHRP 22-15 Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes N/A Accident year X N/A State where the accident occurred X N/A County where the accident occurred X N/A Impact angle (longitudinal axis of vehicle and primary axis of feature) X 1 N/A Separation angle (longitudinal axis of vehicle at last contact and primary axis of feature) X 1,2 N/A Vehicle yawing angle at impact (between longitudinal axis of vehicle and direction of travel) X 1 N/A Vehicle rotation at impact (about vertical axis) X N/A Impact speed (based on vehicle/barrier deformation) X N/A Post-impact vehicle trajectory (qualitative) X 2 Notes: 1- Assumes horizontal accelerometers in perpendicular directions 2- Requires extension of recording to post-crash time period The reader should refer to the appendices for a complete listing and classification of the variables recommended by the NCHRP 22-15 research program. 3.4.11 Trucks Involved in Fatal Accidents (TIFA) The TIFA database combines information from the FARS database pertaining to accidents involving medium and heavy trucks (GVWR > 10,000 lbs) with more detailed information about the involved truck and operating authority [Matteson and Blower, 2003]. Maintained by the University of Michigan Transportation Research Institute (UMTRI) since 1980, this comprehensive database consists of approximately 250 elements.

71 Table 3-14 presents the results of the examination of the TIFA database and subsequent extraction of potential EDR data elements. Table 3-14. TIFA Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes AccidentFactor1 First factor related to accident X 5 AccidentFactor2 Second factor related to accident X 5 AccidentFactor3 Third factor related to accident X 5 AccidentHour Hour in which accident occurred (hh) X AccidentMinute Minute in which accident occurred (mm) X AirBag Air bag availability/function X AvoidType Crash avoidance maneuver X 5 Axles Number of axles X 1 BodyType Vehicle body type X 1 City GSA geographical city code X County County in which accident occurred X DayofMonth Day of month in which accident occurred X DayofWeek Day of week in which accident occurred X DriverFactor1 First driver factor related to accident X 5 DriverFactor2 Second driver factor related to accident X 5 DriverFactor3 Third driver factor related to accident X 5 DriverFactor4 Fourth driver factor related to accident 5 DriverPresent Driver presence X X 2 DriverWeight Driver weight X 2 HoursDriving Hours driving X InitialImpact Initial impact (clock direction) X 6 JulianDate Accident date - Julian X LightCondition Light condition X 3 Maneuver Vehicle maneuver X 5 MannerCollision Manner of collision X ModelYear Vehicle model year X 1 Month Month in which accident occurred X OccupantType Type of occupant X 2 PowerUnitType Power unit type X 1 PwrUnitAxles Power unit, number of axles X 1 PwrUnitMake Power unit, make X 1 PwrUnitYear Power unit, year X 1 Restraint Restraint system use X Rollover Rollover X SeatingPos Seating position of occupant in accident X 2 Speed Estimated travel speed at time of accident X State State in which accident occurred X StateName State name in which accident occurred X StraightBodyStyle Straight truck, body style X 1 SurfaceCondition Roadway surface condition 5 TruckModel Truck model X 1 TruckType Type of truck (using VIN series) X 1 VIN Vehicle identification number X 1 VINLength Length of the vehicle identification number X 1 VehicleFactor1 First vehicle factor related to accident X X 5 VehicleFactor2 Second vehicle factor related to accident X X 5

72 Variable Name Variable Description Current Technology Future Technology Notes VehicleMake Vehicle make X 1 VehicleModel Vehicle model X 1 Weather Weather condition X 4 WeightClass Vehicle weight (using VIN series), by weight class X 1 Year Year in which accident occurred X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- Assumes a weight sensor in all vehicle seating positions. 3- May be inferred from usage of lights 4- May be inferred from usage of windshield wipers 5- May be inferred from precrash information from ABS systems or steering or braking information 6- Assumes horizontal accelerometers in perpendicular directions The reader should refer to the appendices for a complete listing and classification of the variables contained in the TIFA Database. 3.4.12 Motor Carrier Management Information System (MCMIS) – Crash File Operated and maintained by the Federal Motor Carrier Safety Administration (FMCSA), the MCMIS crash file contains information from state police reports pertaining to crashes involving drivers and vehicles of motor carriers [FMCSA, 2004]. The database was started in 1993 and contains approximately 100 data elements for every federally recorded crash involving a motor carrier. Table 3-15 presents the results of the examination of the MCMIS database and subsequent extraction of potential EDR data elements. Table 3-15. MCMIS Extracted Data Elements Variable Name Variable Description Current Technology Future Technology Notes Acdtcnty The 3-digit worldwide geographical code for the county in which the crash occurred. X Acdtdate The date on which the crash occurred. X Acdtmun The name of the municipality (city or township) in which the crash occurred. X Acdtmuncd The 5-digit code for the municipality (city or township) in which the crash occurred as implemented by FIPS PUB 55-2. X Acdttime The military time at which the crash occurred. X Cmlvehicax The number of axles, including auxiliary axles, under the motor vehicle. Axles include all common axis of rotation of one or more wheels, whether power driven or freely rotating. X 1 Light Light condition at the time and place of the crash. X 2 Month Month of crash X

73 Variable Name Variable Description Current Technology Future Technology Notes Rdsurf The condition of the road surface at the time and location of the crash. X 4 State State abbreviation in which crash occurred X StateName State name in which crash occurred X Truckbus Indication of whether the vehicle involved in the crash was a truck (t) or bus (b). X Vehicgvwr Weight rating of the power unit of the vehicle. X 1 Vehicidno Vehicle Identification Number (VIN) of the motor vehicle. X 1 Weather The predominant weather condition at the time and place of the crash. X 3 Year Year of crash X Notes: 1- Can be derived from VIN and associated Vehicle Parameters Database 2- May be inferred from usage of lights 3- May be inferred from usage of windshield wipers 4- May be inferred from precrash information from ABS systems or steering or braking information The reader should refer to the appendices for a complete listing and classification of the variables contained in the MCMIS Database. 3.4.13 Accident Database Needs vs. EDR Data Element Availability One important use of EDR data will be replace or improve data collection for the accident databases. The research team has methodically examined eight existing crash databases and three recommended database formats for candidate EDR data element needs. As shown in Table 3-16, EDR data promises to significantly improve the efficiency of database collection for existing accident statistic databases. Table 3-16. Accident Database Needs vs. EDR Data Element Availability Variables which could be provided by EDR data Database or Database Format Number of Database Variables Current Technology Future Technology FARS 175 10 45 NASS/CDS 400 23 39 NASS/GES 130 8 19 HSIS 120 5 14 LBSS 250 13 27 MMUCC 75 5 19 VEHDB crash tests 180 11 15 NCHRP 350 test 100 6 24 NCHRP 22-15 60 1 8 TIFA 250 15 37 MCMIS 100 - 16

74 3.5 Summary of Data Elements Which Could Be Collected by EDRs A universal catalog was assembled containing the potential EDR data elements extracted from the literature review and the target databases. EDR elements not currently collected for the target databases but representing a possible contribution to roadside safety have been identified and integrated into this listing. Table 3-17 presents the results of the assimilation of extracted data elements. Table 3-17. Catalog of Database Elements Which Could Be Collected by EDRs Variable Description EDR Element(s) Notes Date of accident Crash Date Time of accident Crash Time Location of the Crash (includes latitude, longitude, state, county, and city or municipality) Crash Location (GPS Coordinates) 4 Number of recorded events in accident Event counter Manner of collision PDOF Heading angle for this vehicle Vehicle Direction / Heading Air bag deployment (Driver, passenger, other) Airbag Deployment Parameters Air bag system failure Airbag Diagnostic Codes Max delta V Longitudinal, Lateral, and Normal acceleration 11 Lateral component of delta V Lateral Acceleration 11 Longitudinal component of delta V Longitudinal Acceleration 11 Impact speed Vehicle Speed Impact angle X,Y,Z acceleration and yaw, pitch, and roll rates 3 Vehicle yaw angle at impact Yaw Rate Vehicle separation angle X,Y,Z acceleration and yaw, pitch, and roll rates 3 Vehicle exit speed (i.e. redirectional crash) Vehicle Speed Vehicle stability (before, after and during event) Yaw, roll, and pitch rates 5 Vehicle pitch rate Pitch Rate Vehicle roll rate Roll Rate Vehicle yaw rate Yaw Rate Direction of initial roll Roll Rate vs. time Rollover event indication Roll Rate vs. time Vehicle trajectory (before, after and during event) x,y,z Acceleration, Yaw, Pitch, Roll rates 2 Tire inflation pressure Indicator - Tire inflation pressure Vehicle mileage Vehicle mileage Attempted avoidance maneuver Pre-crash steering angle, braking, engine speed, throttle position, stability control system status, and accelerator position 6 Roadway surface condition Anti-lock braking and traction control systems 7 Light conditions Headlights, crash date/time 8

75 Variable Description EDR Element(s) Notes Acceleration (X,Y,Z directions) Acceleration (X,Y,Z directions) Atmospheric conditions Windshield Wipers, Outside Temperature 9 Vehicle in emergency use at the time of the crash? Siren Indicator Number of occupants within vehicle Occupant Classification Sensor or Number of Occupants Sensor Occupant location in the vehicle Occupant Classification Sensor Occupant weight Occupant Classification Sensor 10 Manual belt system use Belt Status Principal direction of force PDOF Vehicle identification number VIN Antilock brakes Indicator – Antilock brakes Vehicle body type VIN 1 Vehicle curb weight VIN 1 Front/rear wheel drive VIN 1 Driver presence in vehicle Occupant Classification Sensor Four wheel drive VIN 1 Vehicle make VIN Vehicle model VIN Vehicle model year VIN Type of vehicle VIN 1 Restraint type Belt Status, Airbag parameters Wheelbase (original) VIN 1 Test vehicle engine displacement (liters) VIN 1 Maximum width of the vehicle VIN 1 Gross vehicle weight rating VIN 1 Point of Impact (on vehicle) PDOF Indicates the type of sensor EDR Model Location of sensor on vehicle EDR Model Axis direction for sensors measuring vector quantities EDR Model Time increment of the measurement (microseconds) Time Stamp Data status (validity) EDR diagnostic codes, VEDI status, Event Recording Complete Power unit and Engine Specifications (heavy trucks) VIN 1 Occupant Impact Velocity (X-direction) Longitudinal Acceleration Occupant Impact Velocity (Y-direction) Lateral Acceleration Theoretical Head Impact Velocity Longitudinal and Lateral Acceleration Ridedown acceleration (X-direction) Longitudinal Acceleration Ridedown acceleration (Y-direction) Lateral Acceleration Post-Impact Head Deceleration x,y,z Acceleration Acceleration Severity Index x,y,z Acceleration Post-impact max vehicle roll angle Roll Angle / Roll Rate Post-impact max vehicle pitch angle Pitch Rate Post-impact max vehicle yaw angle Yaw Rate Date and Time of accident notification ACN - Date and Time of accident notification Roadway Departure Lateral Lane Position Sensor

76 Notes 1. Vehicle Identification Number + Vehicle Parameter Database 2. The vehicle trajectory can be reconstructed knowing the (a) vehicle acceleration on the longitudinal, lateral, and vertical axes (b) vehicle yaw, pitch, and roll rates, and (c) final resting position/orientation of the vehicle (obtained from site inspection). For planar collisions without rollover, trajectory can be obtained from a more limited set of parameters: (a) acceleration on the longitudinal and lateral axes (b) the vehicle yaw rate, and (c) final resting position/orientation of the vehicle (obtained from site inspection). 3. Computed from the vehicle trajectory. In the case of a collision with a roadside feature, the location and orientation of the fixed object must also be measured from site inspection. 4. County, city, municipality, or state can be obtained by use of GPS coordinates combined with a Geographic Information System (GIS). 5. Assumes that vehicle stability can be inferred from yaw, pitch, and roll rates 6. Assumes that crash avoidance maneuvers can be inferred from pre-crash braking, engine speed, steering angle, and accelerator pedal position. 7. Assumes that roadway surface condition can be inferred from anti-lock braking and traction control systems time history or activation. 8. Light conditions can be inferred from headlights, the time of day, and date. 9. The possibility of rain can be inferred from the use of windshield wipers. 10. Assuming occupant size determined using weight sensors. 11. Delta-V versus time can be computed by integrating acceleration versus time.

77 3.6 Prioritization of EDR Data Elements for Roadside Safety Analysis This section presents a prioritization of EDR data elements by their importance to roadside safety analysis. Earlier sections of this report have presented (1) existing and potential EDR data elements, and (2) Roadside Safety data needs which could be fulfilled by EDRs. The following section describes the results of an exercise in which the research team drew upon the in-depth expertise of subject specialists in roadside safety from the AASHTO Technical Committee on Roadside Safety (formerly the AASHTO Task Force for Roadside Safety) to prioritize EDR data elements by their importance to roadside safety. 3.6.1 Approach On September 25, 2002, the PI met with the AASHTO Technical Committee on Roadside Safety in St. Louis, Missouri, and made a presentation entitled “NCHRP 17-24, Use of Event Data Recorder (EDR) Technology for Roadside Crash Data Analysis: A Status Report”. The PI followed the presentation with a breakout group exercise designed to prioritize EDR Data Elements for Roadside Safety Analysis. The Task Force participants were broken into nine teams. Each team was composed of three to four persons. To facilitate the discussion, each breakout group was assigned one of the following three hypothetical case studies and asked to consider the use of EDR data to analyze the case: Scenario 1. Revision of NCHRP 350. In the year 2015, NCHRP 350 is revised yet again. Evaluate the ‘relevancy’ of NCHRP 350 test requirements. 2. Super-SUVs. In the year 2020, Super-SUVs (Ford Excursions and Chevy Surburbans) have become extraordinarily popular and now account for one-half of the fleet. Evaluate how roadside safety hardware performs with this massive new segment of the vehicle fleet. 3. Accidents on Rural Roads. Evaluate the safety problem of two-lane rural roads and ditches. The participants were each given the table of 37 EDR data elements shown in Table 3-18. The table was assembled from existing or proposed EDR data elements compiled from several different automakers. The participants were told that, because of limited memory in a hypothetical future EDR design, only ten of these elements could be recorded in the event of a crash. They were asked to discuss their scenario and pick the 10 elements which would be most useful in analyzing the case. Finally, the participants were asked to rank order each element by its importance to the analysis. A score of 10 was to be

78 assigned to the most important of the 10 data elements, a score of 1 was to be assigned to the least important of the 10 elements, and a blank score assigned to any unranked elements. The teams were also encouraged to suggest additional elements which were not on the list of EDR data elements. At the completion of the exercise, a quick synopsis of the results was presented to the group for discussion. A more complete analysis of the results is presented below. 3.6.2 Results The results of the prioritization exercise are presented in Table 3-19. Eight of nine teams completed the exercise: three Scenario 1 teams, two Scenario 2 teams, and three Scenario 3 teams. As shown in Table 3-19, the scores assigned to each element were summed first by all teams assigned to a scenario and then for all teams. The maximum score possible was 80 (10 times 8 teams). In addition to the list of given EDR data elements, the combined teams suggested six other data elements, listed at the bottom of Table 3-19, which would be useful for the evaluation of roadside safety. Table 3-20 presents the prioritization of the data elements based on the combined rank ordering of the breakout group participants. The highest priority element was vehicle speed. Second highest priority was yaw rate presumably chosen to determine if the vehicle was tracking upon impact. Third highest priority was the crash location which, given an accurate geographical inventory of roadside safety devices, could be used to correlate the crash with the type of device struck. These top ranked data elements were remarkably consistent across all scenario groups. With few exceptions, the teams independently assigned highest priority to the same top ten data elements despite the fact that the teams were analyzing three very different scenarios. Knowledge of the pre-crash configuration of the striking vehicle was considered paramount. Knowledge of the impact loads as measured by delta-V or crash pulse was a close second in priority. The participant responses were not completely uniform of course. While the participants analyzing Scenario 1 (NCHRP 350 revisions) and Scenario 3 (Rural Road Accidents) were very consistent, participants analyzing Scenario 2 (Super-SUVs) indicated the need for even more in-depth knowledge pre- impact configuration. 3.6.3 Findings This section has described the results of an exercise in which the collective judgment of subject experts in roadside safety from the AASHTO Technical Committee on Roadside Safety was systematically captured to prioritize EDR data elements by their importance to roadside safety. The collective judgment of the AASHTO group indicated that: • Data elements that measure the pre-crash configuration of the vehicle are very important to the analysis of roadside safety. Four of the top ten data elements measured the pre-impact configuration including vehicle speed, yaw rate, roll rate, and lateral acceleration.

79 • Data elements measuring crash performance of the vehicle-roadside system were also of high priority (4 of the top 10) and included lateral delta-V, longitudinal delta-V, lateral crash pulse, and longitudinal crash pulse. • Crash location is of high priority. Given an accurate geographical inventory of roadside safety devices, crash location could be used to correlate the crash with the type of device struck. • Data elements which measured performance of the occupant restraint system, environmental conditions, or driver performance were considered to be important but not of high priority.

80 Table 3-18. OEM Event Data Recorder Data Elements Category Data Element/Description Priority PreCrash Acceleration - lateral Acceleration - longitudinal Brake Position (% where 0 is no brake and 100 is full brake) Clutch status (position or %) Engine RPM Engine RPM Quality Factor Engine Torque Engine Torque Quality Factor Roll rate Steering wheel angle Throttle position Time from ignition on (measured in # of 15 min intervals) Traction control Transmission / gear selection (PRNDL position) Vehicle speed Yaw Rate Vehicle Crash Crash-pulse lateral Crash-pulse longitudinal Delta-V lateral Delta-V longitudinal Time to Max Delta V Restraint Air bag deployment attempt made (yes/no) Air bag(s) inflation time (from time of crash to time of inflation) Air bag(s) status (Front & Side) Airbag(s) on/off switch position (Suppression system status) Belt pre-tensioner status of each passenger Manual Belt Status (On / Off) Time from algorithm wake up to deployment attempt Environment Global Time and date Location of Crash (Latitude / Longitude) Outside Temperature Identifier EDR Serial Number / Model / Version VIN Ignition Cycles Maintenance Battery system voltage Data Validity Check Diagnostic Codes Other (Please List Below and on back of page)

81 Table 3-19. Results of EDR Data Elements Prioritization Exercise Scenario 1 – Revised NCHRP 350 Scenario 2 – Super SUVs Scenario 3 – Rural Roads w/ Ditches Category Data Element/Description Team1 Team2 Team3 Team4 Team5 Team6 Team7 Team8 Scena rio 1 Score Scena rio 1 Score Scena rio 1 Score Total Score PreCrash Acceleration - lateral 2 3 9 5 5 9 5 19 Acceleration - longitudinal 1 4 8 4 5 8 4 17 Brake Position 7 4 3 8 7 4 11 22 Clutch status (position or %) Engine RPM Engine RPM Quality Factor Engine Torque Engine Torque Quality Factor Roll rate 8 8 3 4 5 3 19 4 8 31 Steering wheel angle 2 5 4 1 2 5 5 12 Throttle position 2 2 4 4 Time from ignition on Traction control Transmission / gear selection (PRNDL position) Vehicle speed 10 10 5 10 10 7 7 9 25 20 23 68 Yaw Rate 9 9 4 6 3 6 8 4 22 9 18 49 Vehicle Crash Crash-pulse lateral 7 3 2 10 7 5 10 22 Crash-pulse longitudinal 8 2 1 9 8 3 9 20 Delta-V lateral 6 9 9 9 6 15 9 15 39 Delta-V longitudinal 7 10 1 8 10 7 17 9 17 43 Time to Max Delta V

82 Results of EDR Data Elements Prioritization Exercise (continued) Scenario 1 – Revised NCHRP 350 Scenario 2 – Super SUVs Scenario 3 – Rural Roads w/ Ditches Category Data Element/Description Team1 Team2 Team3 Team4 Team5 Team6 Team7 Team8 Scena rio 1 Score Scena rio 1 Score Scena rio 1 Score Total Score Restraint Air bag deployment attempt made (yes/no) Air bag(s) inflation time 6 6 6 Air bag(s) status (Front & Side) Airbag(s) on/off switch position Belt pre-tensioner status of each passenger 3 3 3 Manual Belt Status (On / Off) 5 1 2 5 3 8 Time from algorithm wake up to deployment attempt 4 4 4 Environment Global Time and date Location of Crash (Latitude / Longitude) 5 6 7 6 8 6 10 11 13 24 48 Outside Temperature 3 1 4 4 Identifier EDR Serial Number / Model / Version VIN 2 5 1 5 2 5 6 13 Ignition Cycles Maintenance Battery system voltage Data Validity Check Diagnostic Codes Other Compass Heading 7 7 7 Other Elements Suggested by the Groups: Steering Wheel Angle History (to capture evasive action) Tire Air Pressure Axle Weight Cell Phone Signal Headlight / Tail lights on

83 Table 3-20. Summary of Results of the EDR Data Elements Prioritization Exercise Overall Priority Data Element Data Element Category Scenario 1 Rank Scenario 2 Rank Scenario 3 Rank Total Score 1 Vehicle speed Precrash 1 1 2 68 2 Yaw Rate Precrash 2 6 3 49 3 Location of Crash Environment 6 2 1 48 4 Delta-V longitudinal Crash 4 5 4 43 5 Delta-V lateral Crash 5 4 5 39 6 Roll rate Precrash 3 14 9 31 7 Brake Position Precrash 8 13 6 22 8 Crash-pulse lateral Crash 9 10 7 22 9 Crash-pulse long. Crash 7 15 8 20 10 Acceleration - lateral Precrash 11 3 11 19 11 Acceleration – long. Precrash 12 7 13 17 12 VIN Identifier 17 12 10 13 13 Steering wheel angle Precrash 16 11 12 12 14 Manual Belt Status Restraint 13 16 8 15 Compass Heading Precrash 8 7 16 Air bag(s) inflation time Restraint 10 6 17 Outside Temperature Environment 14 4 18 Throttle position Precrash 14 4 19 Time from algorithm wake up to deployment attempt Restraint 15 4 20 Belt pre-tensioner status of each passenger Restraint 15 3

84 Table 3-21. EDR Data Element Priority for Roadside Safety Analysis Overall Priority Data Element Data Element Category Current EDR Technology Future EDR Technology 1 Vehicle speed Precrash X 2 Yaw Rate Precrash X 3 Location of Crash Environment X 4 Delta-V longitudinal Crash X 5 Delta-V lateral Crash X 6 Roll rate Precrash X 7 Brake Position Precrash X 8 Crash-pulse lateral Crash X 9 Crash-pulse long. Crash X 10 Acceleration - lateral Precrash X 11 Acceleration – long. Precrash X 12 VIN Identifier X 13 Steering wheel angle Precrash X 14 Manual Belt Status Restraint X 15 Compass Heading Precrash X 16 Air bag(s) inflation time Restraint X 17 Outside Temperature Environment X 18 Throttle position Precrash X 19 Time from algorithm wake up to deployment attempt Restraint X 20 Belt pre-tensioner status of each passenger Restraint X

85 3.7 Recommended EDR Data Elements As the preceding analysis has shown, the use of EDR technology is a promising method to augment data collection for existing roadside and vehicle accident databases. In addition, EDR technology can support future roadside safety research needs by providing a new source of crash data previously not feasible to collect. Based upon the preceding analysis, we recommend that the catalog of 66 data elements presented in Table 3-22 be included in future EDRs to support highway crash analysis. Each data element in the table is annotated according to whether it is currently stored in a production vehicle EDR, is included in the NHTSA NPRM on EDRs, or is one of the priority roadside safety data elements identified by the AASHTO Technical Committee on Roadside Safety. Table 3-22. Recommended EDR Data Elements for Highway Crash Data Analysis Data Element / Description Current EDR Technology NHTSA NPRM Element Priority Roadside Element Accelerator Pedal (%) x Brake Pedal (%) x Brake Pedal Position (on / off) x x Diagnostic Codes Active When Event Occurred x EDR Model Version x Engine Speed (rpm) x x Engine Throttle (%) x x Event Counter x x Event Recording Complete x x Frontal Air Bag Suppression Switch, Passenger x x Frontal Airbag Warning Lamp Status x x Frontal Airbag, Driver, Time to 1st Stage Deployment x x x Frontal Airbag, Driver, Time to 2nd Stage Deployment x x x Frontal Airbag, Passenger, Time to 1st Stage Deployment x x x Frontal Airbag, Passenger, Time to 2nd Stage Deployment x x x Ignition Cycles @ Event x x Ignition Cycles @ Investigation x x Longitudinal acceleration x x x Lateral acceleration x x x Acceleration time stamp x Occupant Size Classification, Passenger (Adult, non- Adult) x x Pretensioner, Driver, Time to Deployment x x x Pretensioner, Pass, Time to Deployment x x x Seat Belt Status, Driver (buckled / unbuckled) x x x Seat Belt Status, Passenger (buckled / unbuckled) x x x Seat Position, Driver, Seat in Forward Seat Position x x Side Airbag Driver, Time to Deployment x x

86 Data Element / Description Current EDR Technology NHTSA NPRM Element Priority Roadside Element Side Airbag, Passenger, Time to Deployment x x Side Curtain/Tube Driver, Time to Deployment x x Side Curtain/Tube Passenger Time to Deployment x x Time between Events x x Vehicle speed x x x Antilock braking (engaged / non-engaged) x Brake Status, Engine (on / off) Crash Date Crash Location (Latitude and Longitude) x Crash Time Front Wipers Switch Frontal air bag deployment level – driver x Frontal air bag deployment level – right front passenger x Hazard Lights Switch Headlight Switch Indicator Status - Tire Pressure Monitoring System Indicator Status - VEDI Normal Acceleration x Occupant Position, Driver, out of position x Occupant Position, Passenger out of position x Occupant Size Classification, Driver (Adult, Small Adult) x Seat Position, Passenger, Seat in Forward Seat Position x Stability control (on / off / engaged) x Steering Input (steering wheel angle) x x Temperature - Ambient Air x Traction Control Status Vehicle Identification Number x Vehicle Mileage Vehicle Roll Angle x Yaw Rate x Cell Phone (On / Off) Siren Status (On / Off) Lateral Lane Position Number of Occupants Accident Notification – Date and Time Pitch Rate Principal Direction of Force Roll Rate Vehicle Direction / Heading x Notes Following are a few notes of explanation for the preceding data element recommendations:

87 • The Ignition Cycle Count, a data element currently being stored in some EDRs, is recommended as an interim measure of when the crash occurred. Once EDRs begin to store crash date and time, Ignition Cycle Count should no longer be needed. • Most accident database, e.g. FARS or NASS/CDS, store some information about the occupant restraint system as this facet of a highway crash is well known to be indicative of injury outcome. To date however, information about the restraint system has been limited to data elements such as whether an occupant was buckled and whether the airbag deployed. EDRs provide copious information about the performance of the occupant restraint system components, including airbag deployment level, performance of belt pretensioners and occupant position at the time of deployment. These parameters are an example of the new data elements which are crucial to understanding the outcome of a highway crash, but have never before been available to investigators. They are included in the list of recommended EDR data elements. • Note that if it acceleration versus time is available, then it becomes unnecessary to store delta-V vs. time. Delta-v can be computed by integrating acceleration. Discussion Thirty-two, or almost half of these data elements, are already being stored in production vehicle EDRs. Thirty-eight (38) of these elements are defined in the NHTSA NPRM on Event Data Recorders. In the near-term, we recommend adopting the data elements proposed in the NHTSA NPRM and adding the following five priority roadside safety data elements: (1) Vehicle Identification Number, (2) crash location, (3) yaw rate, and (4) roll rate. In the longer term, we recommend that the entire list be included in the data elements which are stored in future EDRs.

88 3.8 Recommendations for EDR Enhancement Current EDR capabilities will need to be enhanced to better support roadside crash analysis. Following is a list of recommendations: • Increase the EDR recording duration. To capture roadside feature crash performance, future EDRs need to record for a greater length of time than is the current practice. Roadside safety analyses require knowledge of not only the pre-crash trajectory, but also the post-crash trajectory. Currently, this data could be obtained if EDRs, such as the GM SDM, stored ‘pre- crash’ parameters such as vehicle velocity for 5 seconds before and after a crash. Likewise, impacts with roadside features such as a guardrail are relatively long events in comparison with vehicle-to-vehicle crashes. To capture the entire vehicle-to- roadside event, it would be useful if the crash pulse could be recorded for a minimum of 300 milliseconds. • Increase the number of events recorded. A crash is frequently characterized by multiple events. For example as shown in Figure 3-1, a car may first inadvertently leave the road and glance off a guard-rail – the first event, careen into the path of an oncoming car – the second event, and finally strike a tree on the opposite side of the highway – the third event. Event 1 - Guardrail Event 2 Airbag Deploys Event 3 - Tree Figure 3-1. Current EDRs may not capture all events in a crash. Most current EDRs are not equipped to record all the events that may occur in a crash. GM EDRs, for example, are capable of capturing two events: a non- deployment event and a deployment event. For some later GM EDR designs, a

89 deployment level event, which occurs after bag deployment, can record over a non-deployment event. Even these newer GM devices however can only capture two events. Some automakers’ EDRs, e.g. the Ford RCM, are only capable of capturing a single event. As the typical event captured is the event that deployed the air bag, any subsequent events may not be recorded even if these events are more harmful. 54% 29% 11% 6% 0% 10% 20% 30% 40% 50% 60% 1 Event 2 Events 3 Events More than 3 Events Number of Events per Vehicle Fr eq ue nc y (% ) Figure 3-2. Events per Vehicle for NASS/CDS 2000-2002 EDR Cases Figure 3-2 presents the distribution of events per vehicle for the 2000-2002 NASS/CDS cases with a successful EDR download (Gabler et al, 2004). 46% of the EDR cases involved two or more events. In 17% of the cases, the vehicle was involved in three or more events. As GM EDRs can only store a maximum of two events, it is likely that potential EDR data were “lost” from one or more events in these crashes. • Expand the Definition of an Event The literature also suggests that the definition of an ‘event’ will be an important design consideration for future EDRs. Currently, an event is a crash. In addition to this type of event, the literature indicates that roadway departure, with or without an impact, is also an important event. Lane-keeping and roadway departure warning systems are now entering the market which could be adapted for this purpose. Accurate recording and retrieval of roadway departure events would be invaluable for encroachment studies.

90 3.9 Conclusions The objective of this chapter was to catalog and prioritize EDR data needs which support vehicle and roadside safety research and design. The specific objectives were to determine the potential of EDR technology (1) to augment data collection for existing roadside and vehicle accident databases, and (2) to support future roadside safety research needs by providing a new source of crash data previously not feasible to collect. Because an EDR is vehicle-mounted, the device is, of course, limited to what can be measured on the vehicle. However, the preceding analysis has shown that many data elements needed for roadside safety research or collected in for accident databases can be provided by EDR technology. Our findings are as follows: • A review of the roadside safety literature suggests that many of the data elements recommended for collection by previous research studies can either be obtained by current EDR devices or in future EDR designs. Examples of critical research data needs are pre-crash vehicle trajectory, post-crash vehicle trajectory, and the orientation of the vehicle (yaw, pitch, roll) at the time of impact. • Many of the data element needs can be obtained if future EDRs begin to store VIN. The VIN contains complete information on the vehicle make, model, year, and curb weight. When these identifiers are combined with a database such as the NHTSA Vehicle Parameter database, the data needs for vehicle geometry can also be obtained. • EDR data has the promise to significantly improve the efficiency of database collection for existing accident statistic databases. Based upon the methodical examination of eight existing crash databases and three recommended database formats, we conclude that a significant fraction of data elements currently being collected could be provided by either existing or future EDR data elements. For example, 55 of the 175 FARS data elements could be provided by EDRs. For state accident databases designed to meet MMUCC format, 24 of the 75 recommended data elements could be provided by EDRs. • The priority of EDR data elements was ranked in an exercise in which the collective judgment of subject experts in roadside safety from the AASHTO Technical Committee on Roadside Safety was systematically captured to prioritize EDR data elements by their importance to roadside safety. The collective judgment of the AASHTO group indicated that data elements that measure the pre-crash orientation of the vehicle was of highest priority to the analysis of roadside safety. Crash location was also deemed to be of high priority. It is interesting to note that six of the ten highest priority data elements can be collected by current EDR technology. Only four of the highest priority data elements will require enhancements to existing EDR technology

91 Our recommendations are: • Based on a comparison of EDR capabilities and highway crash data analysis needs, a catalog of 66 recommended data elements has been developed. Nearly half of these data elements, are already being stored in production vehicle EDRs. Thirty-eight (38) of these elements are defined in the NHTSA NPRM on Event Data Recorders. In the near-term, we recommend adopting the data elements proposed in the NHTSA NPRM and adding the following four priority roadside safety data elements: (1) crash location, (2) Vehicle Identification Number, (3) yaw rate, and (4) roll rate. In the longer term, we recommend that automakers store the entire list of data elements in future EDRs. • To capture roadside feature crash performance, future EDRs need to record for a greater length of time than is the current practice. Roadside safety analyses require knowledge of not only the pre-crash trajectory, but also the post-crash trajectory. Currently, this data could be obtained if EDRs, such as the GM SDM, stored ‘pre- crash’ parameters such as vehicle velocity for 5 seconds before and after a crash. Likewise, impacts with roadside features such as a guardrail are relatively long events in comparison with vehicle-to-vehicle crashes. To capture the entire vehicle-to- roadside event, it would be useful if the crash pulse could be recorded for a minimum of 300 milliseconds. • EDRs need to record an increased number of events. EDRs which record only a single event, e.g. the current Ford design, lose approximately one-half of the events. EDRS which record only two events, e.g. the current GM design, lose approximately 17% of the events. An EDR which records 3 events, on the other hand, would capture 94% of the crash events. • The definition of an ‘event’ should be expanded to include roadway departures. Currently, an event is a crash. In addition to this type of event, roadway departure, with or without an impact, is also an important event. Lane-keeping and roadway departure warning systems are now entering the market which could be adapted for this purpose. Accurate recording and retrieval of roadway departure events would be invaluable for encroachment studies.

92 3.10 References Bligh, Roger P. “Performance of Current Safety Hardware for NCHRP 350 Vehicles”. Transportation Research Circular #440, TRB, National Research Council, April 1995, pp. 29-34. Council, Forrest, M. and Stewart, J. Richard. “Attempt to Define Relationship between Forces to Crash-Test Vehicles and Occupant Injury in Similar Real-World Crashes”. Transportation Research Record 1419, Transportation Research Board, Washington, D.C., 1993. Eskandarian, A., Bahouth, G., Digges, K., Godrick, D., and Bronstad, M. Improving the Compatibility of Vehicles and Roadside Safety Hardware. Preliminary Draft Final Report, NCHRP Project 22-15, Transportation Research Board, October 2002. Erinle, Olugbenga. An Analysis of Guardrail and Median Barrier Accidents Using the Longitudinal Barrier Special Studies (LBSS) File, Volume II: User’s Guide. U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-92-099, August 1994. FHWA, Highway Safety Information System: Guidebook for the Utah State Data Files, Volume I: SAS File Formats, U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-01-56, March 2000. http://www.hsisinfo.org/pdf/01-056.pdf FHWA, Highway Safety Information System: Guidebook for the California State Data Files, Volume I: SAS File Formats, U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-00-137, March 2000. http://www.hsisinfo.org/pdf/00- 137.pdf FHWA, Highway Safety Information System: Guidebook for the Maine State Data Files, Volume I: SAS File Formats, U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-00-082, March 2000. http://www.hsisinfo.org/pdf/00- 082.pdf FHWA, Highway Safety Information System: Guidebook for the Minnesota State Data Files, Volume I: SAS File Formats, U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-01-058, March 2000. http://www.hsisinfo.org/pdf/01- 058.pdf FHWA, Highway Safety Information System: Guidebook for the Michigan State Data Files, Volume I: SAS File Formats, U.S. Department of Transportation, Federal Highway Administration, FHWA-RD-01-118, January 2001. http://www.hsisinfo.org/pdf/MIvol1.pdf

93 FMCSA, Motor Carrier Management Information System (MCMIS) Data Dissemination Catalog & Documentation, Federal Motor Carrier Safety Administration, http://mcmiscatalog.fmcsa.dot.gov/beta/Catalogs&Documentation/default.asp, (viewed December 2004) Gabler, H.C., Hampton, C.E., and Hinch, J., “Crash Severity: A Comparison of Event Data Recorder Measurements with Accident Reconstruction Estimates”, SAE Paper 2004-01-1194 (2004) Gabauer, D.J., and Gabler, H.C., “Evaluation of Acceleration Severity Index Threshold Values Utilizing Event Data Recorder Technology”, Proceedings of the TRB 84th Annual Meeting, Paper 05-2220, Washington, DC (January 2005) Gabauer, D.J., and Gabler, H.C., “A Methodology to Evaluate the Flail Space Model Utilizing Event Data Recorder Technology,” Transportation Research Record: Journal of the Transportation Research Board, No. 1890, TRB, National Research Council, Washington, DC (2004) Glennon, J.C. and Wilton, C.J. Effectiveness of Roadside Safety Improvements Vol. I – A Methodology for Determining the Safety Effectiveness of Improvements on All Classes of Highways. FHWA-RD-75-23. United States Department of Transportation, Federal Highway Administration, Washington, D.C., 1974. Hall, J.W., Turner, D.S., and Hall, L.E. “Concerns About Use of Severity Indexes in Roadside Safety Evaluations”. Transportation Research Record 1468, National Research Council, Washington, DC, Dec, 1994, pp 54-59. Hunter, W.W. and Council, F.M., “Future of Real World Roadside Safety Data”, Roadside Safety Issues Revisited, Transportation Research Circular 453, TRB, National Research Council, February 1996, pp. 38-54. Information Systems and Services, Inc.(ISSI) NHTSA Test Reference Guide, Volume I: Vehicle Tests, National Highway Traffic Safety Administration, Version 5, May 2001. http://www-nrd.nhtsa.dot.gov/pdf/software/vehdb_v5.pdf Kullgren, A., Lie, A., and Tingvall, C., “Crash Pulse Recorder (CPR) – Validation in Full Scale Crash Tests”, Accident Analysis and Prevention, Vol. 27, No. 5 (1995) Kullgren, A., Ydenius, A, and Tingvall, C. “Frontal Impacts with Small Partial Overlap: Real Life Data from Crash Recorders”, International Journal of Crashworthiness, Vol. 3, No. 4 (1998) Kullgren, A., Krafft, M., Nygren, A, and Tingvall, C. “Neck Injuries in Frontal Impacts: Influence of Crash Pulse Characteristics on Injury Risk”, Accident Analysis and Prevention, v.32, no. 2 (2000)

94 Mak, K. K., “Problems Associated with Police-Level Accident Data in Evaluation of Roadside Appurtenance Performance,” Transportation Research Circular No. 256, Transportation Research Board, Washington, D.C., 1983. Mak, K. K., Sicking, D. L., and Ross, H. E., Jr., “Real-World Impact Conditions for Ran- Off-the-Road Accidents,” Transportation Research Record 1065, Transportation Research Board, Washington, D.C., 1986. Mak, K.K. and Sicking, D.L. Development of Roadside Safety Data Collection Plan, FHWA-RD-92-113, Texas Transportation Institute, Texas A&M University System, College Station, Texas, February 1994. Mak, K.K. “Methods for Analyzing the Cost-Effectiveness of Roadside Safety Features”, Roadside Safety Issues, Transportation Research Circular 435, TRB, National Research Council, January 1995, pp. 42-62. Mak, King K. and Bligh, Roger P. Recovery-Area Distance Relationships for Highway Roadsides: Phase I Report. NCHRP Project 17-11, Transportation Research Board, January 1996. Mak, King K., Bligh, Roger P., and Griffin, Lindsay I. Improvement of the Procedures for the Safety Performance Evaluation of Roadside Features. NCHRP Project 22-14 Final Report, Transportation Research Board, November 2000. Matteson, Anne and Blower, Daniel. Trucks Involved in Fatal Accidents: Codebook 2000. Report UMTRI-2003-07. University of Michigan Transportation Research Institute. US Department of Transportation. March 2003. McCullough, C.A., Hollowell, W.T., and Battisti, P. “Applications of NHTSA’s Vehicle Parameter Database”, Society of Automotive Engineers Paper No. 950360 (1995) McGinnis, Richard G., and Swindler, Kathleen M. “Roadside Safety in the 21st Century”. Proceedings of the Conference on Traffic Congestion and Traffic Safety in the 21st Century, ASCE Highway Division and Urban Transportation Division, Jun 8-11 1997, Chicago, IL, pp 118-124. Michie, J.D. “Roadside Safety: Areas of Future Focus”, Roadside Safety Issues Revisited, Transportation Research Circular 453, TRB, National Research Council, February 1996, pp. 30-37. NHTSA, Model Minimum Uniform Crash Criteria (MMUCC). US Department of Transportation, National Highway Traffic Safety Administration, Federal Highway Administration, DOT-HS-808-745, August 1998. http://www-nrd.nhtsa.dot.gov/edr- site/uploads/MMUCCaugust98.pdf

95 NHTSA, National Automotive Sampling System (NASS) General Estimates System (GES) Analytical User's Manual: 1988-2001. U.S. Department of Transportation, National Highway Traffic Safety Administration. http://www-nrd.nhtsa.dot.gov/pdf/nrd- 30/NCSA/GES/01GESAUM.pdf NHTSA, 1988 - 1996 NASS CDS Variable-Attribute Structure Manual. U.S. Department of Transportation, National Highway Traffic Safety Administration, February 1998. http://www-nass.nhtsa.dot.gov/NASS/MANUALS/CDS8896.pdf NHTSA, National Automotive Sampling System (NASS) Crashworthiness Data System (CDS) Analytical User's Manual: 2000 File. U.S. Department of Transportation, National Highway Traffic Safety Administration. http://www- nass.nhtsa.dot.gov/NASS/cds/AnalyticalManuals/aman2000.pdf (2001) Opiela, Kenneth S. and McGinnis, Richard M. “Strategies for Improving Roadside Safety”. 1998 Transportation Conference Proceedings. Powers, R., Dearsaugh, W., et al. “Breakout Group Discussion B: Selection and Design of Roadside Safety Treatments”, Roadside Safety Issues, Transportation Research Circular 435, TRB, National Research Council, January 1995, pp. 73-75. Powers, R., “Breakout Group Discussion E: In-Service Evaluation and Barrier Performance Data Research Needs”, Roadside Safety Issues Revisited, Transportation Research Circular 453, TRB, National Research Council, February 1996, pp. 114-115. Ray, M. H., J. D. Michie, and M. Hargrave. “Events That Produce Occupant Injury in Longitudinal Barrier Accidents”. Transportation Research Record 1065, TRB, National Research Council, Washington, D.C., 1986, pp. 70-75. Ray, Malcolm H. “Safety Advisor: Framework for Performing Roadside Safety Assessments”. Transportation Research Record 1468, National Research Council, Washington, D.C., December 1994, pp 34-40. Ray, Malcolm H., Carney, John F., and Opiela, Kenneth S. “Emerging Roadside Safety Issues”. TR News #177, National Research Council, Washington, D.C., Mar-Apr 1995, pp 32-35. Ray, Malcolm H., Hargrave, Martin W., Carney, John F. III, and Hiranmayee, K. “Side Impact Crash Test and Evaluation Criteria for Roadside Safety Hardware”. Transportation Research Record 1647, National Research Council, Washington, D.C., November 1998, pp 97-103. Ross, H.E., Jr., Perera, H.S., Sicking, D.L., and Bligh, R.P., Roadside Safety Design for Small Vehicles, NCHRP Report 318, Transportation Research Board, Washington, D.C., 1988.

96 Ross, H. E., Jr., Sicking, D. L., Zimmer, R. A., and Michie, J.D., Recommended Procedures for the Safety Performance Evaluation of Highway Features, NCHRP Report 350, Transportation Research Board, Washington, D.C., 1993. Tessmer, Joseph M., Fatality Analysis Reporting System (FARS) Analytic Reference Guide 1975-1999, U.S. Department of Transportation, National Highway Traffic Safety Administration (1999) Viner, John G., Council, Forest M., and Stewart, J. Richard. “Frequency and Severity of Crashes Involving Roadside Safety Hardware by Vehicle Type”. Transportation Research Record 1468, National Research Council, Washington, D.C., December 1994, pp 10-18. Viner, John G., “The Roadside Safety Problem”. Roadside Safety Issues, Transportation Research Circular 435, TRB, National Research Council, January 1995, pp. 17-29. Viner, John G. “Risk of Rollover in Ran-Off-Road Crashes”. Transportation Research Record 1500, National Research Council, Washington, D.C., July 1995, pp 112-118.

Next: 4. EDR Retrieval and Archival Methods: Current Methods, Limitations, and Issues »
Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Web Only Document 75: Use of Event Data Recorder (EDR) Technology for Highway Crash Data Analysis examines the legal issues surrounding EDRs and the consumer acceptability of EDR data collection.

  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!