et al. designed one of the first comprehensive motorcycle tests where a rider was ejected into the barrier to examine guardrail designs. Test conditions were chosen to best replicate real-world crashes based on available data (Quincy et al. 1988). Expanding upon this work, Koch and Brendicke (1989) performed a crash test to examine guardrail impacts from motorcyclists. Compared to earlier studies, the impact angle and impact speed were lowered to precisely investigate which portion of the guardrail causes injury when impacted by the rider (Koch and Brendicke 1989). In 2007, Peldschus et al. summarized multiple studies that conducted motorcycle-barrier tests (Peldschus et al. 2007). This includes a follow-on study by Quincy in 1998 at an increased impact speed and two different orientations of the anthropomorphic test device (ATD) (Quincy 1998). An ATD is a tool commonly used in crash testing to measure potential human injury in different testing conditions. Gärtner et al. (2006) completed two upright impact tests with shallower impact angles and a higher speed to test head and thorax injury response to impacts with barriers. These test conditions were chosen to replicate common higher speed impacts that occur on the roadway. These tests have resulted in the implementation of European Standards Committee Technical Specification (CEN TS) 1317-8 as the standard test for motorcycle protection systems on barriers in Europe. Similar to many previous studies, CEN TS 1317-8 prescribes a 30-degree impact angle with the ATD striking the barrier head first. The standard prescribes both a 60-km/h and a 70-km/h impact. One of the most recent studies was conducted by the Texas A&M Transportation Institute (TTI) with similar testing conditions to evaluate motorcyclist injury prevention additions on current barriers (Silvestri Dobrovolny et al. 2019). TTI motorcycle testing conditions were specified by the Texas Department of Transportation and were chosen to represent real-world barrier impacts (Table 59).

Crash and Encroachment Studies

The foundation for barrier test impact conditions is data from real-world encroachments (i.e., the vehicle leaves the traveled way) and real-world encroachment crashes. The current test conditions in MASH are based on the crash encroachments collected in the NCHRP 17-22 database. The NCHRP 17-22 dataset extracted encroachment data from the NASS/CDS case years 1997 to 2000 and a few cases from 2004 (Mak et al. 2010). NASS/CDS is an in-depth, nationally representative dataset of crashes involving at least one passenger vehicle that was towed from the scene due to damage and contains detailed information on the environment conditions, the vehicle damage, and occupant injuries (NCSA 2015). Despite the in-depth information, NASS/CDS captures very little information on the roadside. The NCHRP 17-22 dataset used the scaled NASS/CDS scene diagrams to measure the encroachment angles, lateral extent, and road geometry. Based on the vehicle deformation, the impact speed was reconstructed by the NCHRP 17-22 research team. As a follow-on study to NCHRP 17-22, the NCHRP 17-43 dataset continued extracting roadside encroachment data from more recent case years (Riexinger and Gabler 2019). The NCHRP 17-43 database contains 1,581 roadside encroachments extracted from NASS/CDS case years 2011 to 2015.

Most encroachment studies, including NCHRP 17-22 and NCHRP 17-43, have focused on passenger vehicles, while relatively few data collection studies have been conducted for motorcycles in the U.S. In contrast, there have been many data collection efforts focused on motorcycles in Europe and New Zealand [Association of European Motorcycle Manufacturers (ACEM) 2004; Stefan et al. 2003]. The first motorcycle in-depth crash data collection effort in the U.S. was collected from 1976 to 1979 in Los Angeles (Hurt et al. 1981). This dataset included an on-scene crash investigation, an interview with the motorcyclist, and injury data. While this dataset is not specific to roadside encroachments, Ouellet (1982) used the Hurt dataset to identify hazards for motorcyclists. Ouellet discussed that road and roadside designs focused on passenger vehicles may increase the risk of crashes for motorcyclists. A more recent data collection effort was conducted by Gabler et al. (2022) in the NCHRP 22-26 project from 2010 to 2016. This project identified and collected in-depth data on 21 single-vehicle motorcycle-barrier crashes where a rider was admitted into a Level 1 trauma center. In 2016, FHWA released the new MCCS database. This database comprises 351 injury crashes, of which 82 cases are single-vehicle motorcycle crashes that occurred in California.

motorcycle crashes in California where the motorcyclist was transported to a Level 1 trauma center.

NCHRP 22-26: NCHRP project 22-26 collected in-depth data for crashes involving a motorcyclist who struck a roadside barrier and was transported to a Level 1 trauma center.

Case Selection Criteria

The NCHRP 17-88 database contains 1,581 crashes incorporated from NCHRP 17-43, 124 crashes incorporated from LTCCS, 47 crashes from MCCS, and 21 from NCHRP 22-26. These in-depth crash datasets contain motorcycles, passenger cars, utility vehicles, vans, light trucks (pickups), single-unit trucks, and tractor-trailer trucks. Because the dataset is comprised of cases from multiple data sources, not every data element was available for every case. Some additional data elements such as speed limit and object struck are being coded for the NCHRP 17-43 cases. Therefore, the case count was provided for each analyzed variable in Table 61. Partial departure crashes, where the subject vehicle impacted a roadside object without the vehicle center of gravity departing the roadway, were excluded from the analysis. When analyzing impact events, such as the object struck or the impact angle, only the first impact was considered. The subsequent impacts were affected by the first impact, and that could lead to very unusual impact dynamics. For example, a vehicle that strikes a narrow object may yaw significantly due to the impact before striking a second object. In this analysis, when a vehicle splits into multiple components, only the component that originally contained the driver was considered. For example, if a tractor-trailer truck was separated into the trailer and the cab, only the trajectory of the cab was considered. Additionally, if a motorcyclist was separated from the bike during a crash, the trajectory of the motorcyclist was considered.

Encroachment and Impact Condition Characterization

Six vehicle types were considered: passenger cars, vans, light trucks, motorcycles, single-unit trucks, and tractor-trailer trucks. These vehicle categories are based on the terminology NHTSA uses in the source NASS/CDS dataset. Passenger cars include sedans, convertibles, and station

wagons. Vans include mini-vans and van-based light trucks. Utility vehicles include compact utility vehicles (CUVs) and sport utility vehicles (SUVs). Light trucks include pickup trucks and other trucks with a gross vehicle weight rating under 10,000 lb. NCHRP 17-43 data were merged with NASS/CDS to evaluate case weight in the context of passenger vehicles. Using case weights assigned in NASS/CDS allows frequency estimates to be nationally representative (Radja 2016). The analysis of cases from LTCCS was unweighted, and MCCS does not include any case weights. The cases from NCHRP 22-26 were combined with MCCS to increase the number of motorcycle crashes available for analysis. In this analysis, the departure angle, posted speed, impact angle, road type, and object struck were compared between the vehicle types. These comparisons help to understand the differences in roadside impacts involving motorcycles due to their unique characteristics.

5.1.3 Results

Impact and Departure Angles

For each type of vehicle, the departure angles on each side of the road were combined and plotted using a cumulative distribution function (CDF). All included departure cases represent full road departures. Departures from both sides of the road occurred between 0 and 180 degrees, with 0 degrees being tangent to the roadway in the direction of travel and 180 degrees being tangent to the roadway in the opposite direction.

In general, the distributions of departure angles corresponding to each vehicle type were nearly identical up to 13 degrees, or roughly 65% (Figure 16). The single-unit truck 85^th percentile departure angle was 33 degrees, which was higher than the other vehicle types (Figure 16). The distribution of motorcycle and tractor-trailer truck departure angles were very similar to each of the passenger vehicle types. The median motorcycle departure angle was 27 degrees. The largest motorcycle departure angles were much larger than those of passenger vehicles due to a control loss prior to the departure.

side, and a separated motorcyclist is completely removed from the motorcycle. Over 75% of recorded motorcycle crashes included multiple events. During the first event, almost 80% of cases were upright. By the second event, almost 50% of cases were separated. This indicates that a large percentage of riders lose contact with the motorcycle during the first event and are separated during any subsequent events (Figure 19).

The first roadside crash event for passenger vehicles typically involved fixed objects such as trees and poles. In contrast, the most common first roadside crash event for single-unit and tractor-trailer trucks involved concrete barriers, guardrails, or a rollover. For motorcycles, the most common objects contacted were curbs and the guardrail length of need (Figure 20). Motorcycles likely have a larger percentage of recorded impacts with curbs due to their reduced size and lack of encompassing body. Curb impacts are likely not severe enough to be recorded as an event by the original crash investigator for larger vehicle types such as passenger vehicles and large trucks. Motorcycle impacts with guardrails are overrepresented in this dataset because of the sampling criteria of the NCHRP 22-26 project.

occurred on a curve. Furthermore, motorcycle crashes occurred on roads with a smaller radius of curvature (Figure 24).

Injury Outcomes

Injury outcomes were typically more severe for motorcycles and large trucks compared to passenger vehicles, with a dramatic difference in motorcycles. When a crash occurs, a KABCO injury designation is given by police on scene, while Abbreviated Injury Scale (AIS) injury scores

are assigned by medical professionals after triage. AIS levels are not recorded in LTCCS. On the AIS scale, motorcycle injuries are more common at the most severe designations of Critical and Maximum (Figure 25). Motorcycles have increased injury due to the lack of safety features that prevent injury in passenger vehicles like airbags and seatbelts. When comparing light trucks and passenger vehicles using KABCO designations, they are similar at the two most severe levels of incapacitating injury and fatal. However, light trucks have a much greater percentage of injuries that are not incapacitating, while passenger vehicles have a greater percentage of no or possible injury (Figure 26). Light trucks experience a higher frequency of high-speed limit rollover crashes, which can increase injury severity.

5.1.4 Discussion

Objects Struck

When studying objects struck, the percentages are relatively consistent across the different vehicle types, with a larger percentage of passenger vehicles striking poles and trees, motorcycles striking curbs, and a larger percentage of large trucks striking concrete barriers. The increase in curbs hit by motorcycles is likely due to the small body size and lack of an encompassing frame on motorcycles. Impacts with curbs are likely not severe enough to be recorded in standard impacts

in large trucks or passenger vehicles. The higher number of concrete barriers hit by large trucks compared to other vehicle types could be due in part to large trucks primarily operating on higher speed limit multilane roads. In contrast, passenger vehicles were more likely to hit fixed objects such as poles and trees, likely due to passenger vehicles traveling on lower speed limit, single-lane roads. High speed limit, multilane roads are likely interstates with a designated clear zone and a greatly increased prevalence of concrete barriers. Single-lane, low speed limit roadways have the potential to be much closer to fixed objects, increasing the frequency that vehicles collide with said objects.

Barrier Impact Tests

The motorcycle impact angles were very similar to passenger cars and light trucks, which are the current MASH test vehicles (Table 63). The passenger car and light truck 85^th percentile impact angles were very similar to the MASH test criteria. The 85^th percentile impact angle for large trucks, both single-unit and tractor-trailer trucks, was higher than the MASH test conditions. This is likely due to LTCCS sampling higher severity crashes.

Previous studies have performed motorcycle impacts in upright and separated crash conditions because motorcyclists were often ejected from their motorcycle. This study found that during the first impact, the majority of motorcycles were upright. In only 20% of these first impacts, either the motorcycle was down with the rider still on the motorcycle or the rider had separated from the motorcycle. To date, this down but coupled with the rider configuration has not been performed with physical testing. After the first impact, the majority of riders have been separated from their motorcycle. After three impacts, no motorcycles remained upright. Future motorcycle-barrier crash tests should alter the configuration of the motorcycle and the motorcyclist depending on the intended installation type. If the barrier is likely to be the first impact once installed, then an impact with the motorcycle in the upright position would be the most representative. However, if the barrier is likely to be the second impact, such as an installation behind a curb, then an impact test with the motorcyclist separated from the motorcycle would be most representative.

The MASH testing impact angle for sedans and pickup trucks was very close to the 85^th percentile impact angle in the NCHRP 17-88 database. If MASH were to adopt a similar practical worst-case scenario (85^th percentile) for motorcycle crashes, the impact angle would be 24 degrees. This is lower than the 30-degree impact angle in CEN 1317 part 8, the currently adopted standard in Europe. Based on the motorcycle impacts in the NCHRP 17-88 database, the majority of riders are upright on the motorcycle just prior to the first impact event. MASH could prioritize tests with an

upright configuration to most closely match the roadside impacts observed in the U.S. This would be different from head-first, prone position of the rider alone during the impact test prescribed in CEN 13-17 part 8. However, this test would be very similar to the upright test conducted at TTI (Silvestri Dobrovolny et al. 2019).

Limitations

The in-depth data from MCCS were combined with NCHRP 22-26 to increase the sample size of motorcycle impacts in the NCHRP 17-88 database. While this is all of the available in-depth motorcycle crash data since 2000, it is important to note the small sample of motorcycle crashes analyzed. The results from these motorcycle barrier crashes may not generalize outside of the collection area (California and North Carolina). However, until additional motorcycle crash data are collected, this study represents the largest analysis of motorcycle impact conditions.

5.1.5 Conclusions

This study analyzed the largest collection of in-depth motorcycle crash data and compared the impact conditions of passenger vehicles, motorcycles, and large trucks. Tractor-trailer trucks consistently had lower departure and impact angles compared to passenger vehicles, while motorcycles tended to have very similar departure and impact angles compared to passenger vehicles. The higher center of mass in large trucks and the decreased stability of motorcycles resulted in a much higher frequency of rollover in the sampled road departure crashes. Additionally, it was found that motorcyclists are typically upright during the first impact event. By the second impact event, the majority of motorcyclists were separated from the motorcycle. Passenger vehicle and large truck testing standards are determined by the 85^th percentile of impact angles in the U.S. Following this precedent, future motorcycle-barrier impact studies should consider a 24-degree impact angle. Additionally, the vast majority of motorcycle crashes in the NCHRP 17-88 database were upright during the first impact. Testing should consider testing in the upright configuration at a 24-degree impact angle to most closely reflect current crashes occurring on the roadway.