Motorcycle crashes potentially preventable by ...

Motorcycle crashes potentially preventable by passenger vehicle crash avoidance technology August 2017

Eric R. Teoh Insurance Institute for Highway Safety

ABSTRACT Objective: To identify and quantify the motorcycle crash population that would be potential beneficiaries of the crash avoidance technology available on passenger vehicles. Methods: Two-vehicle crashes between a motorcycle and a passenger vehicle that occurred during 201115 were classified by type, with consideration to functionality of three classes of passenger vehicle crash avoidance technology: front crash prevention, lane maintenance, and blind spot detection. Results were expressed as the percentage of crashes potentially preventable by each type of technology, based on all known types of two-vehicle crashes and based on all crashes involving motorcycles. Results: Front crash prevention had the largest potential to prevent two-vehicle motorcycle crashes with passenger vehicles: 4 percent of fatal crashes and 13 percent of police-reportable crashes of any severity. The three technologies in sum had the potential to prevent 10 percent of fatal two-vehicle crashes and 23 percent of police-reportable crashes. However, since two-vehicle crashes with a passenger vehicle represent fewer than half of all motorcycle crashes, these technologies represent a potential to avoid 4 percent of all fatal motorcycle crashes and 10 percent of all police-reported motorcycle crashes. Discussion: Refining the ability of passenger vehicle crash avoidance systems to detect motorcycles represents an opportunity to improve motorcycle safety. Expanding the capabilities of these technologies represents an even greater opportunity. However, even fully realizing these opportunities can affect only a minority of motorcycle crashes and does not change the need for other motorcycle safety countermeasures such as helmets, universal helmet laws, and antilock braking systems. Keywords: Motorcycle crashes, crash avoidance technology, crash type, motorcycle safety, highway safety, forward collision warning, automatic emergency braking, blind spot detection, lane departure warning, lane-keeping assist





2

INTRODUCTION Motorcycles present substantial challenges in highway safety. They lack structures necessary to protect their occupants in crashes, and their small size makes them harder for other road users to see. Moreover, it is especially challenging to judge the speed of oncoming motorcycles because of their narrow width and because they typically have only one headlamp. Per mile traveled, motorcyclists are nearly 5 times more likely than passenger car occupants to be injured in traffic crashes and nearly 29 times more likely to be killed (NHTSA, 2017). Efforts to improve motorcycle safety largely have focused on helmets and helmet use laws, which have been shown many times over to reduce the risk of dying in crashes (Liu et al, 2009; Houston and Richardson, 2008). However, helmets don’t prevent all deaths, and many motorcycle crashes are injurious regardless of helmet use, so crash prevention also is crucial. Antilock braking systems help motorcyclists avoid many crashes (Teoh, 2013; Basch et al., 2015) but do not directly address the problem that motorcycles are not as visible to other road users. Rider training courses typically teach riders to position themselves in more visible locations, but this doesn’t change the physical characteristics of motorcycles, limiting potential benefits. The proliferation of crash avoidance technologies on passenger vehicles represents an opportunity to reduce the conspicuity problem motorcyclists face. These technologies, which exist under a wide variety of brand names and functionality specifics, use advanced sensors to monitor some aspects of the driving environment and warn the driver or intervene when they detect a possible collision. There are three primary classes of such systems that potentially could prevent crashes in which a passenger vehicle strikes a motorcycle – front crash prevention, lane maintenance, and blind spot detection. Front crash prevention systems use radar, laser, or video camera sensors (or some combination thereof) to monitor the road environment ahead. Forward collision warning (FCW) systems warn the driver if a collision is likely, and automatic emergency braking (AEB) systems apply the vehicle’s service brakes to prevent or mitigate a forward impact if the driver does not take corrective action. A vehicle may be equipped with FCW, AEB, or both, but typically AEB systems include FCW. These systems focus on vehicles traveling in the same direction, and are thus most relevant to rear-end crashes. Among current systems, there is much variation in whether AEB functions at all speeds and whether it brings the vehicle to a complete stop, whether FCW/AEB detects stationary vehicles, and in whether these systems detect smaller road users like motorcyclists, bicyclists, pedestrians, or animals. Front crash prevention systems, especially those with both FCW and AEB, have been shown to reduce rear-end crash rates in the passenger vehicle fleet (Cicchino, 2017a). As pointed out by Cicchino (2017a), these systems have been increasingly offered as optional or standard equipment in new passenger vehicles, and 20 automakers that represent 99 percent of the US automobile market pledged to make FCW and AEB standard equipment on virtually all new passenger vehicles by September 2022. Lane maintenance systems use video cameras to identify lane lines and warn the driver if an unintentional lane departure is detected. Some systems intervene by applying steering input or braking wheels on one side to deter the vehicle from departing the lane if the driver does not correctively steer. Systems recognize intentional



3

lane changes when the driver engages the turn signal. These systems function only at higher speeds, typically above about 40 mph. Some systems allow the driver to adjust the sensitivity to lane deviations to account for driving habits. The research on lane departure systems is less clear than it is for forward collision systems. While lane departure warning systems have shown benefits of 11 percent for the crashes they were designed to address (Cicchino, 2017b), they were not associated with decreases in overall collision insurance losses (HLDI, 2016a; HLDI, 2016b). One factor related to effectiveness is that drivers may find lane maintenance systems more annoying than front crash prevention, and they may be more likely to turn them off (Eichelberger and McCartt, 2014; Reagan and McCartt, 2016; Reagan et al., 2017). Blind spot detection systems use radar, ultrasonic, or video camera sensors to detect other vehicles present in the equipped vehicle’s blind spot. Systems typically illuminate a warning light on the corresponding side mirror or A-pillar whenever a vehicle is present, and some systems issue an additional warning if they detect that the driver intends to change lanes in that direction when an adjacent vehicle is present. Some systems intervene to try to prevent lane changes when a vehicle is detected in the blind spot. Blind spot systems have been shown to reduce lane-change crashes by 14 percent (Cicchino, 2017c), and property damage liability insurance claim rates by 2-11 percent (HLDI, 2012; HLDI, 2016c). These crash avoidance systems are constantly improving in terms of sensor and software technology, detection capability, and human factors. Systems have become more likely to be designed to intervene by braking or steering as hardware and software advance, improving detection and minimizing false positives. As they improve, it is important to refine their ability to detect all road users, including motorcycles. Some systems claim to detect motorcycles, some claim not to, and many do not address this directly. Lane maintenance need not detect motorcycles to deliver a benefit. The purpose of this paper is to quantify how many motorcycle crashes potentially could be affected if these systems on passenger vehicles detect motorcycles. METHODS Data were extracted from the Fatality Analysis Reporting System (FARS) and the National Automotive Sampling System - General Estimates System (NASS-GES) for crashes involving motorcycles during the five-year period 2011-15. Motorcycle crashes were classified as single-vehicle, two-vehicle with passenger vehicle, twovehicle other, and 3+ vehicle. The analysis focused on two-vehicle crashes between a motorcycle and a passenger vehicle. Passenger vehicle crash avoidance technology could be expected to affect crashes with 3+ vehicles, but it is difficult to determine the events of such crashes reliably. Two-vehicle crashes were classified by type and by relevance to the three crash avoidance technologies on passenger vehicles. The relevance of crashes to the three technologies was coded primarily using the crash type variable in FARS and NASS-GES (acc_type), which is coded at the vehicle level and considers the movements of both vehicles prior to impact. A coding diagram (NHTSA, 2016) for this variable is provided in Appendix A. In crash types in which impact direction was uncertain from this variable alone, the initial impact point variable was considered as



4

well. For example, crashes in which a motorcycle, traveling in the same direction, turns across the path of a passenger vehicle would be potentially relevant to front crash prevention if the passenger vehicle struck the motorcycle, as opposed to the motorcycle striking the side of the passenger vehicle. So, relevant crashes of this type were restricted to those in which the passenger vehicle’s initial point of impact was coded as 11-, 12-, or 1o’clock. Similarly, the variable “pre-event movement (prior to recognition of critical event)” was used to exclude or classify certain situations, such as when the passenger vehicle was stopped or was intentionally changing lanes. For example, a same-direction sideswipe crash in which the passenger vehicle was coded as “changing lanes” or “overtaking” is unlikely to be relevant to lane maintenance but likely relevant to blind spot detection. Lane maintenance systems tend to function only at speeds of about 40 mph or higher, so crashes were considered relevant only if the speed limit was 35 mph or higher (or on an interstate highway in the case of same-direction sideswipe). Appendix B provides definitions of all crash situations considered in this study, and further information about these variables is available in the 2015 FARS/NASS-GES Coding and Validation Manual (NHTSA, 2016). NASSGES includes statistically imputed values for missing data among several of these variables; these were used whenever available. Two-vehicle crashes between motorcycles and passenger vehicles were tabulated by severity and type, then classified as to their relevance to each crash avoidance technology. Technologies’ relevance to crashes were expressed as counts and as percentages of known crash type. Since no crash type was considered relevant to more than one technology, the total number of crashes addressable by the three technologies was taken as the sum of the counts for each technology. Percentages also were calculated based on all motorcycle crashes, not just two-vehicle, to express the potential benefit to motorcycle safety in general. RESULTS Table 1 outlines motorcycle crashes during 2011-15 by number of vehicles and severity. Motorcycle crashes, 86 percent of which were injurious or fatal, largely fell into two categories regardless of severity: singlevehicle or two-vehicle where the other vehicle was a passenger vehicle. Crashes involving three or more vehicles and two-vehicle crashes with a vehicle other than a passenger vehicle (e.g. large truck, another motorcycle, unknown) were far less frequent. Two-vehicle motorcycle crashes, where the other vehicle was a passenger vehicle, accounted for almost half of all motorcycle crashes. These crashes are further broken down by crash types in Table 2. Percentages were calculated on the basis of known crash types, with denominators of 7,973 fatal crashes, 144,009 nonfatal injury crashes, and 192,163 crashes of any severity reported to police. Crashes relevant to front crash prevention mostly were rear-end crashes in which a passenger vehicle rear-ended a motorcycle. These accounted for a larger share of police-reported crashes (12 percent of known types) than fatal crashes (4 percent). Other crashes relevant to front crash prevention included those involving a frontal impact on the passenger vehicle and where a motorcycle turned into the path of a passenger vehicle in the same direction, or where a motorcycle moving in the same



5

direction turned across the path of a passenger vehicle, but these were infrequent, accounting for fewer than 1 percent of crashes. Overall, front crash prevention was relevant to 349 fatal crashes (4 percent), 14,905 nonfatal injury crashes (10 percent), and 24,593 police-reported crashes (13 percent) over the five-year study period. Crashes relevant to lane maintenance systems accounted for a similar proportion of fatal crashes (4 percent), nonfatal injury crashes (3 percent), and all police-reported crashes (4 percent) of known type. However, head-on crashes in which the passenger vehicle unintentionally departed its lane were far more common in fatal crashes – 205 fatal crashes (3 percent) vs. 175 police-reported crashes (