Real-world effects of rear cross-traffic alert on police-reported backing crashes

Transcription

1 Real-world effects of rear cross-traffic alert on police-reported backing crashes February 2018 Jessica B. Cicchino Insurance Institute for Highway Safety

2 ABSTRACT Objective: Several parking assistance systems are available to help drivers back up. Rear crosstraffic alert warns drivers when they are reversing and cross-traffic approaches the rear of their vehicle, as might happen when backing out of a perpendicular parking space in a lot. The goal of this study was to examine the effect of rear cross-traffic alert on backing crashes. Methods: Negative binomial regression was used to compare police-reported backing crash rates in 25 U.S. states per insured vehicle year between General Motors and Mazda vehicles with rear crosstraffic alert and the same vehicle models without the optional system, controlling for the presence of other parking assistance systems and additional factors that may affect crash risk. Results: Crash involvement rates in backing crashes overall were 22% lower among vehicles with rear cross-traffic alert than among vehicles without the system when averaged between the two manufacturers. Rates were 32% lower among vehicles with the system than without in two-vehicle backing crashes where the vehicles were traveling in perpendicular directions. Both reductions were statistically significant. Conclusions: Rearview cameras and rear automatic braking, which are designed to help drivers detect and brake for obstacles directly behind the vehicle, have been shown in previous research to be effective in reducing backing crashes reported to the police. Rear cross-traffic alert can complement these systems by preventing potential backing crashes that other parking assistance systems are not designed to detect. Manufacturers should equip vehicles with rear cross-traffic alert, in addition to other parking assistance systems like rear automatic braking, to maximize the number of backing crashes prevented. Keywords: backing assist; collision avoidance; crash warning; parking assist; parking crashes; rear crosspath detection; 1

3 1. Introduction Backing can be a challenging task for drivers because visibility behind vehicles is limited (Kidd & Brethwaite, 2014; Mazzae & Barickman, 2009). Approximately 188,000 passenger vehicles were backing in crashes reported to the police in the United States in 2015, according to data from the National Highway Traffic Safety Administration ([NHTSA], 2016), making up 2% of all passenger vehicle crash involvements. Because NHTSA s data only include crashes on public roadways and therefore exclude backing crashes occurring in parking lots and on private driveways, this figure underestimates the true number of vehicles involved in these crashes (Perez et al., 2011). To help drivers avoid backing crashes, a number of parking assistance systems have been developed that increase rear visibility or provide alerts or emergency braking when an obstacle is detected behind the vehicle while reversing. Some parking assistance systems are proving to be effective at reducing backing crashes in the real world. Cicchino (2017a) compared the police-reported crash experience of Honda, Mazda, Mercedes- Benz, and Subaru vehicles with an optional rearview camera to that of the same vehicles where the rearview camera was not purchased. On average, backing crash involvement rates were 17% lower among vehicles with a rearview camera than among those without. The effects of a rearview camera combined with rear parking sensors, which alert the driver to obstacles behind the vehicle, on police-reported backing crashes among General Motors (GM) vehicles were examined by Flannagan, Kiefer, Bao, Leblanc, and Geisler (2014) and Cicchino (2018). Both found benefits for the combination of systems in excess of 40%. Vehicles with these two GM backing systems as well as a rear automatic braking system, which provides emergency braking when an object is detected behind the vehicle and the driver does not respond to warnings, had backing crash involvement rates that were 78% lower than vehicles with none of the systems (Cicchino, 2018). In New Zealand, the odds that a pedestrian strike was the result of a vehicle backing was 41% lower among vehicles with a standard rearview camera, 31% lower among vehicles with standard rear parking sensors that provided warnings, and 30% lower among vehicles with both systems standard when compared with vehicles where the systems were optional or not available (Keall, Fildes, & Newstead, 2017). 2

4 Rearview cameras and rear parking sensors are designed to improve visibility and object detection directly behind the vehicle, but the range of these technologies typically does not extend considerably to areas on either side of the vehicle s rear. A different type of parking assistance system, rear cross-traffic alert, uses radar to detect approaching vehicles that might cross the path of a backing vehicle. The system can be useful to drivers whose rear view may be obscured, for example, when backing from a parking space between two larger vehicles. Since first becoming available in the United States on the model year 2009 Dodge Grand Caravan and Chrysler Town & Country, in model year 2018 rear cross-traffic alert is available on 54% of new vehicle series, although most often as an optional (43%) rather than standard (11%) feature (Highway Loss Data Institute [HLDI], 2018). Not much is known about the real-world effects of rear cross-traffic alert on crashes. HLDI (2017) compared rates of property damage liability insurance claims, which cover damage an at-fault vehicle causes to other vehicles and property, between Mazda vehicles with rear cross-traffic alert and the same make and model vehicles without the system, controlling for other collision avoidance systems on the vehicle and other factors that may affect crash risk, and found no significant difference in claim rates among vehicles with and without the system. However, HLDI did not have information on the circumstances surrounding crashes and was not able to restrict their analyses to the backing crashes that are relevant to rear cross-traffic alert. The goal of the current study was to examine the effect of rear cross-traffic alert on the type of crashes the system was designed to prevent. Police-reported backing crash involvement rates among GM and Mazda vehicles with optional rear cross-traffic alert were compared to rates among the same vehicle models without the technology, controlling for other parking assistance systems on the vehicles and other factors that may affect backing crash risk. 2. Methods 2.1 Vehicles Study vehicles included GM and Mazda series where rear cross-traffic alert was offered as an optional feature (Table 1). Vehicle identification numbers (VINs) of GM and Mazda vehicles with rear 3

5 cross-traffic alert and other parking assistance systems were provided by the manufacturers; this information was used to identify the features equipped on vehicles in the sample. The study was limited to vehicles from these two manufacturers, because information was not available from other manufacturers on the presence or absence of rear cross-traffic alert and other parking assistance systems at the VIN level. Table 1. Study vehicle series and model years Make Series Model years Buick LaCrosse 4D 2WD Buick LaCrosse 4D 4WD Buick Regal 4D 2WD Buick Regal 4D 4WD Cadillac ATS 2D 2WD 2015 Cadillac ATS 2D 4WD 2015 Cadillac ATS 4D 2WD Cadillac ATS 4D 4WD Cadillac CTS 4D 2WD Cadillac CTS 4D 4WD Cadillac Escalade 4D 2WD 2015 Cadillac Escalade 4D 4WD 2015 Cadillac Escalade ESV 4D 2WD 2015 Cadillac Escalade ESV 4D 4WD 2015 Cadillac SRX 4D 2WD Cadillac XTS 4D 2WD Cadillac XTS 4D 4WD Chevrolet Impala 4D Chevrolet Suburban 1/2T 4D 2WD 2015 Chevrolet Suburban 1/2T 4D 4WD 2015 Chevrolet Tahoe 4D 2WD 2015 Chevrolet Tahoe 4D 4WD 2015 GMC Yukon 4D 2WD 2015 GMC Yukon 4D 4WD 2015 GMC Yukon XL 1/2T 4D 2WD 2015 GMC Yukon XL 1/2T 4D 4WD 2015 Mazda 3 4D Mazda 3 5D Mazda 6 4D 2WD Mazda CX-3 2WD/4WD 2016 Mazda CX-5 4D 2WD 2016 Mazda CX-5 4D 4WD 2016 Mazda CX-9 4D 2WD Mazda CX-9 4D 4WD Mazda MX-5 Miata Convertible D=two-door, 4D=four-door, 5D=five-door, 2WD=two-wheel drive, 4WD=four-wheel drive 4

6 The rear cross-traffic alert systems from both manufacturers use radar to detect other objects or vehicles approaching the rear of the vehicle from the side when in reverse. GM s system sounds three (left or right) beeps or produces three (left or right) pulses in the Safety Alert Seat (i.e., seat vibrations) that will sound or vibrate on the side of the vehicle from which the detected vehicle is approaching. A red triangle with a left- or right-pointing arrow also appears on the left or right side of the Rear Vision Camera (rearview camera) screen, depending on the direction of the detected vehicle. Mazda vehicles with rear cross-traffic alert give an auditory warning and display a light in the appropriate side view mirror when an approaching vehicle or object is detected. If the vehicle has a Rearview Monitor (rearview camera), a warning indication also appears in the camera display. Parking assistance systems other than rear cross-traffic alert were offered on study vehicles. Relationships among systems that were offered are illustrated in Table 2. All GM vehicles with rear cross-traffic alert also had both the Rear Vision Camera and Rear Parking Assist (rear parking sensors); the combination of the Rear Vision Camera and Rear Parking Assist was also offered on vehicles without rear cross-traffic alert. Automatic Parking Assist, which helps drivers with parallel and perpendicular parking by finding appropriately-sized parking spaces and providing automated steering into them, was offered on GM vehicles with and without rear cross-traffic alert. Only GM vehicles without rear crosstraffic alert may have had Rear Parking Assist alone (i.e., without other parking assistance systems), the Rear Vision Camera alone, or no parking assistance system; and only vehicles with rear cross-traffic alert may have had Rear Automatic Braking, which provides automatic emergency braking while backing if drivers do not respond to alerts when an obstacle is detected directly behind the vehicle. On Mazda vehicles, the Rear View Monitor and rear cross-traffic alert were offered both together and individually, so that a vehicle may have had the Rear View Monitor only, rear cross-traffic alert only, both systems, or neither system. These other parking assistance systems were controlled for in analyses. For both manufacturers, vehicles with rear cross-traffic alert had other parking assistance systems more often than vehicles without it. 5

7 Table 2. Parking assistance systems offered on vehicles with and without rear cross-traffic alert (RCTA). GM With RCTA All Some Some Without RCTA Some Some Some Some Some Mazda With RCTA Some Some Without RCTA Some Some Note: All=all vehicles from the manufacturer with/without RCTA also have the system named in the column; Some=some vehicles from the manufacturer with/without RCTA also have the system named in the column Crash data No other parking assistance system Police-reported crash data were obtained from 25 states where backing vehicles could be identified and where VINs of individual vehicles involved in crashes could be obtained, so that crash data could be merged with vehicle feature data. States and years during which data were obtained included Indiana and Rhode Island during ; New Mexico during ; Delaware, Florida, Georgia, Iowa, Kansas, Louisiana, Michigan, Minnesota, Mississippi, Missouri, Nebraska, New Jersey, Oklahoma, Pennsylvania, South Dakota, Tennessee, Texas, Utah, Washington, Wisconsin, and Wyoming during ; and Maryland during Data from 2012 were not included in Mazda analyses, because the oldest Mazda vehicles in the sample were from model year 2014 and thus had no exposure during calendar year Rearview camera (no parking sensors) Backing crashes were defined as crashes where the subject vehicle was backing prior to the crash. Two-vehicle perpendicular backing crashes were considered to be most relevant to rear cross-traffic alert and were defined as crashes involving two vehicles moving in perpendicular directions prior to the crash, where the subject vehicle was backing prior to the crash. Direction of travel was unavailable in three states (Louisiana, Mississippi, and Oklahoma), and so these states were excluded from analyses of twovehicle perpendicular backing crashes. In the 22 states where the variable was available, direction of travel was missing for 6% of crash involvements. Rear parking sensors (no rearview camera) Rearview camera + rear parking sensors Rear automatic braking Automatic parking assist 6

8 Location of the crash was examined to determine if the crash occurred off of the roadway, in a parking lot, and/or on private property. Three states coded for one of these characteristics, 15 coded for two, and 7 coded for all three. Study vehicles were involved in crashes 32,782 times in the 25 study states and 30,337 times in the subset of 22 states where direction of travel was available Exposure data Data on vehicle exposure and characteristics of the rated driver (age, gender, marital status, and insurance risk level) and of the vehicle s garaging location (density of registered vehicles in the ZIP code where the vehicle is garaged) were obtained from HLDI. HLDI s data include approximately 85% of U.S. insured passenger vehicles. Vehicle exposure was expressed in insured vehicle days, and is presented as insured vehicle years in tables. Vehicle feature data, crash data, and insurance exposure data were merged by matching VINs within states; because VINs were matched within states, crashes that occurred in a different state than where a vehicle was insured were not captured. Vehicles that did not appear in the HLDI database or that were insured in a different state than where they crashed were excluded from both the numerator and denominator of crash rates Regression models Negative binomial regression was used to model crash involvement rates per insured vehicle year. Four regression models were constructed: one examining overall backing crash involvement rates and a second examining two-vehicle perpendicular backing crash involvement rates, with each carried out separately for Mazda and GM vehicles. Models controlled for rated driver age (15 34, 35 54, 55 69, 70+, unknown), gender, marital status, and insurance risk level (standard or nonstandard); state; calendar year; and density of registered vehicles (0 499, 500+) in the ZIP code where the vehicle is garaged. Nonstandard risk policies are those under which the policyholder is rated as a high insurance risk; vehicles with these policies file in excess of 30% more collision claims, which cover damage at-fault drivers cause to their vehicles, per insured vehicle year compared with those with standard risk (HLDI, 2004). 7

9 A single variable capturing the vehicle series and model year was included to prevent confounding of vehicle feature effects with other vehicle design changes that may occur between vehicle series and model years. A binary variable was included, indicating the presence or absence of rear crosstraffic alert in all models. Additional variables were included in GM s models, indicating the presence or absence of the combination of the Rear Vision Camera and Rear Parking Assist, Rear Parking Assist alone (i.e., without a Rear Vision Camera), the Rear Vision Camera alone (i.e., without Rear Parking Assist), Rear Automatic Braking, and Automatic Parking Assist. Mazda s models included an additional variable denoting the presence or absence of the Rear View Monitor. Models used a logarithmic link function. Suppose C i represents the number of crash involvements, E i represents exposure (i.e., insured vehicle days), and R i represents the presence or absence of rear cross-traffic alert for vehicle i. Assuming C i is a negative binomial random variable with mean E(C i)=e iλ i, the statistical models were specified as log λ i = log(e(c i/e i)) = β 0 + β 1(R i) + β 2(covariates). Thus, 100(exp (β 1) 1) represented the percent difference in the crash involvement rate for vehicles with rear cross-traffic alert compared with vehicles without Pooled estimates Effect estimates for rear cross-traffic alert and for driver demographic characteristics common across analyses (rated driver age, gender, marital status, and insurance risk level; registered vehicle density in ZIP code where vehicle is garaged) from regression models were pooled between manufacturers using methods outlined below, which are similar to those used in evaluations of the effectiveness of other collision avoidance technologies (Cicchino, 2017a, 2017b; Fildes et al., 2015). Because effects were pooled from only two manufacturers, fixed effects models were used. A weight was assigned to each estimate as follows: ww ii = 1 vv ii where vv ii represents the estimate s variance. Pooled effects were calculated as follows: 8

10 yy = gg ii=1 gg ii=1 ww ii where yy ii is each effect estimate, ww ii is each estimate s weight, and gg is the total number of estimates (in this case, g=2). Since weights were primarily based on the inverse variance of the respective estimates, estimates with high variance contributed less to the pooled effect than estimates with low variance. 3. Results Ninety-five percent confidence intervals (CIs) were computed using the following equation: 95% CCCC = gg ii=1 ww iiyy ii gg ii=1 ww ii ww iiyy ii ± gg ii=1 ww ii Study vehicles were involved in police-reported backing crashes 1,044 times in the 25 study states, making up 3% of all police-reported crash involvements among study vehicles. There were 236 two-vehicle perpendicular backing crash involvements in the 22 states where such crashes could be identified, which comprised 24% of the 997 police-reported backing crash involvements among study vehicles in these 22 states. Among all 1,044 backing crashes, at least 64% occurred off the roadway, in a parking lot, or on private property; 25% involved a parked vehicle; and 70% involved another vehicle that was not parked. Few involved injuries to anyone in the crash (4%) or to a pedestrian or bicyclist (2%). Backing crash involvement rates and two-vehicle perpendicular backing crash involvement rates were lower among vehicles with rear cross-traffic alert than among vehicles without for both manufacturers (Table 3). Demographic characteristics of rated drivers and vehicle garaging locations of vehicles with and without rear cross-traffic alert, weighted by insured vehicle years, are summarized in Table 4. Characteristics were fairly similar between groups. Rated drivers of vehicles with rear cross-traffic alert were more likely to be married than other drivers, and were slightly more likely to be middle-aged and have their vehicle garaged in a less-dense area. 9

11 Table 3. Police-reported backing crash involvement rates in all backing crashes in 25 states and in two-vehicle backing crashes where vehicles were traveling in perpendicular directions in 22 states among study vehicles with rear cross-traffic alert (RCTA) and other parking assistance systems. All backing crashes Two-vehicle perpendicular backing crashes Manufacturer Parking assistance systems Insured vehicle years Crashes Rate (x1000) Insured vehicle years Crashes Rate (x1000) GM None 39, , No RCTA, with other parking assist systems 156, , RCTA + other parking assist systems 283, , Mazda None 29, , No RCTA, with rearview camera 18, , RCTA, no rearview camera 14, , RCTA + rearview camera 129, , Total 672, ,

12 Table 4. Percent of insured vehicle years among vehicles with and without rear cross-traffic alert (RCTA) with various rated driver and vehicle garaging characteristics. Characteristic Without RCTA With RCTA Age Unknown 7 7 Gender Male Female Unknown Marital status Single Married Unknown Insurance risk Standard Nonstandard 5 4 Registered vehicles per square mile in ZIP code where vehicle is garaged Results of negative binomial regression models pooled between the two manufacturers examining the effects of rear cross-traffic alert and other factors on backing crash involvement rates are summarized in Table 5. On average, higher backing crash involvement rates were associated with rated drivers age 70 and older, drivers with nonstandard insurance risk, and vehicles garaged in ZIP codes with a registered vehicle density of 500 or more vehicles per square mile. Rear cross-traffic alert was associated with a significant 22% reduction in backing crash involvement rates and with a significant 32% reduction in two-vehicle perpendicular backing crash involvement rates when effects were pooled between GM and Mazda vehicles. 11

13 Table 5. Pooled effects from negative binomial regression models of backing crash involvement rates. Parameter All backing crash involvements Perpendicular backing crash involvements Estimate Effect % (95% confidence interval) p-value Estimate Effect % (95% confidence interval) Rear cross-traffic alert (ref=without) (-35, -7) (-54, 0) Age (ref=15 34) (-17, 22) (-32, 51) Age (ref=15 34) (-19, 22) (-32, 62) Age 70+ (ref=15 34) (10, 71) (-11, 136) Unknown age (ref=13 24) (-5, 54) (-58, 99) Female (ref=male) (-9, 19) (-27, 28) Unknown gender (ref=male) (-78, -39) (-89, -30) Single (ref=married) (-7, 26) (-21, 47) Unknown marital status (ref=married) (12, 184) (30, 599) Nonstandard insurance risk (ref=standard) (-6, 67) (8, 212) registered vehicles per square mile (ref=0-499) (18, 56) < (18, 116) p-value 12

14 For individual manufacturers (not in table), Mazda vehicles with rear cross-traffic alert had overall backing crash involvement rates that were 34% lower (95% CI: -52%, -10%) than vehicles without and two-vehicle perpendicular backing crash involvement rates that were 49% lower (95% CI: - 73%, -5%); GM s rear cross-traffic alert system was associated with reductions of 15% (95% CI: -32%, 5%) and 19% (95% CI: -50%, 33%), respectively, in these crash types. Effects were statistically significant for Mazda s system and were not for GM s system. However, there was substantial overlap between the confidence intervals associated with effect estimates from the two manufacturers, which indicates that the estimates did not differ significantly between them. 4. Discussion The current study presents the first evidence that rear cross-traffic alert is reducing backing crashes reported to the police. Averaging effects between vehicles from GM and Mazda, backing crash involvement rates were 22% lower among vehicles with rear cross-traffic alert than without. In twovehicle crashes where the backing vehicle struck another vehicle traveling in a direction perpendicular to it, which were considered to be most representative of the types of backing crashes rear cross-traffic alert was designed to prevent, the system reduced crash involvement rates by 32%. The effectiveness of collision warnings depends upon drivers responding appropriately when warnings are received. Two experimental studies suggest that drivers often brake upon hearing rear crosstraffic alerts and avoid encroaching into the path of an approaching vehicle as a result. Seventy drivers recruited by Neurauter, Llaneras, Li, and Green (2011) drove a vehicle either with only a rearview camera or with a rearview camera and rear cross-traffic alert. A vehicle approached the rear of the vehicle in a staged surprise event where drivers backed out of a parking space from between two large SUVs. Prior to this trial, drivers were not aware of either the study purpose or availability of a rear cross-traffic alert. Drivers with wide-view (180 degree field-of-view) and regular-view (130 degree field-of-view) rearview cameras each encroached into the path of the crossing vehicle in 88% of trials, while drivers of vehicles with both a rearview camera and rear cross-traffic alert encroached in 49%, a statistically significant reduction of 44%. Drivers also braked significantly sooner in response to the crossing vehicle in vehicles 13

15 with rear cross-traffic alert than in vehicles without, at 1.2 seconds and 2.1 seconds brake response time, respectively. In a similar experimental setup by Reimer, Mehler, and Coughlin (2010), drivers backed a vehicle with rear cross-traffic alert or a vehicle without it in a parking garage with their visibility obscured by a pole. All drivers using a vehicle with rear cross-traffic alert stopped at an appropriate time to avoid the crossing vehicle, while only 15 of 21 drivers without the technology stopped appropriately. Experimental research suggests that drivers often do not respond appropriately to alerts from traditional rear parking sensors that detect objects directly behind the vehicle. For example, in two studies, one-third to half of drivers who received warnings from rear parking sensors when stationary or moving objects surreptitiously appeared behind their backing vehicle did not respond discernably to the warnings, fewer than half of drivers responded to warnings by applying the brakes, and two-thirds to three-quarters struck the obstacles (Kidd, Hagoski, Tucker, & Chiang, 2015; Llaneras et al., 2005). It is unknown if driver response to rear cross-traffic and traditional parking sensor alerts are meaningfully different, because they have not been compared directly in the same study. However, it is plausible that drivers may be more likely to respond appropriately to rear cross-traffic alerts than to the warnings from traditional rear parking sensors, because rear cross-traffic alerts are more likely to indicate a true hazard and occur when their vision is at least partially obscured while warnings from parking sensors notify drivers of their proximity to objects and true hazards. Expecting that an alert indicates a true hazard increases the chance that a driver chooses to respond (e.g., Bliss, Gilson, & Deaton, 1995). Regardless of how often drivers respond to rear cross-traffic alerts relative to alerts from rear parking sensors, it would still be expected that rear cross-traffic prevention systems that provide emergency braking when the driver does not respond appropriately to alerts would be more effective in preventing backing crashes than rear cross-traffic alerts alone. When it comes to both forward collision prevention (Cicchino, 2017b) and rear parking sensor systems (Cicchino, 2018), those that provide automatic emergency braking, in addition to warnings, have been shown to be more effective at preventing crashes than systems that only warn. Currently available front automatic emergency braking systems are generally designed to detect the rear of other vehicles, but future refinements of these systems 14

16 that respond to cross-traffic configurations could also prevent rear cross-traffic crashes when equipped on the potential striking vehicle. Collision avoidance systems also need to be turned on to be effective. Although it was unknown if rear cross-traffic alert among study vehicles was activated at the time of the crash, owner self-report (Cicchino & McCartt, 2015) and objective observations of the activation status of collision avoidance systems among vehicles taken in for service at dealership service centers (Reagan, Cicchino, Kerfoot, & Weast, 2018) indicate that nearly all owners with rear cross-traffic alert systems keep them turned on. Systems from only two manufacturers were examined, because data were not available on the presence or absence of the systems on vehicles from other manufacturers. Results may not generalize to rear cross-traffic alert systems from other manufacturers with different characteristics. Analyses controlled for factors that affected backing crash risk, such as rated driver age, insurance risk, and the density of registered vehicles in the ZIP code where the vehicle was garaged, but because rear crosstraffic alert was optional equipment on study vehicles, drivers who chose to purchase the technology could differ in uncontrolled ways that may have over or underestimated the true effect of the technology on backing crash rates. Although it could be intuited that drivers who choose to purchase packages that include additional safety equipment may be more safety-conscious than drivers who choose not to, this has not always been the case with other optional safety features. For example, a study linking passenger vehicle and motorcycle insurance claims among owners of motorcycle models where antilock braking systems (ABS) were an optional feature found that motorcycles were slightly more likely to be equipped with ABS as auto insurance claim rates increased, suggesting that riskier drivers were more likely to purchase ABS for their motorcycles (HLDI, 2014). When the effects of ABS on motorcycle insurance claim rates were examined in an analysis similar to the current study, results were nearly identical when an additional control for auto insurance claim rates was added. As shown in Table 2, some GM and Mazda vehicles examined in this analysis offered rear crosstraffic alert on vehicles that were also equipped with additional backing systems. For Mazda, a rearview camera was offered on some vehicles equipped with rear cross-traffic alert. For GM, a rearview camera 15

17 and rear parking sensors were always offered on vehicles with rear cross-traffic alert, and these GM vehicles were sometimes also equipped with rear automatic braking and/or automatic parking assist systems. Hence, despite attempts made in this study to isolate on backing crashes that are uniquely addressed by the rear cross-traffic alert feature, it remains possible that these other backing systems could provide overlapping benefits. Therefore, for vehicles that were also equipped with other backing systems, the reported additional effectiveness of the cross-traffic alert feature may not be as large as that for vehicles not equipped with other backing systems. Backing crashes reported to the police are not representative of all backing crashes that occur. Many states do not require crashes that occur off the public roadway to be reported to the police, but yet, about two-thirds of study crashes occurred off-road, in a parking lot, or on private property. Because some of these crashes were not required to be reported, there may be bias in which crashes ultimately were reported. Also, many property-damage-only backing crashes do not meet states minimum reporting thresholds, but still are worth preventing as they cost time and money. Backing crashes in which the two involved vehicles were traveling perpendicularly were considered to be most relevant to rear cross-traffic alert, but this set likely included some crashes that occurred outside of the systems operating limits, for example, where the reversing vehicle or crossing vehicle was traveling too fast. Several technologies currently exist to assist drivers while backing, including rearview cameras, rear parking sensors that provide proximity alerts, and rear automatic braking. Rear cross-traffic alert addresses a subset of backing crashes that may be missed by other parking assistance systems that monitor the area directly behind the vehicle. Manufacturers should equip vehicles with rear cross-traffic alert, in addition to other highly effective parking assistance systems like rear automatic braking, to maximize the number of backing crashes that can be prevented. While rearview cameras will become standard equipment on most new light vehicles to meet rear visibility requirements in the United States that come into effect May 2018, technologies like rear automatic braking and rear cross-traffic alert will remain optional features on most vehicles on which they are available. Thus, consumers should also be 16

18 made aware that these technologies can be useful additions to the rearview cameras that will already come standard. ACKNOWLEDGEMENTS The author would like to thank David Zuby, David Kidd, and Eric Teoh of the Insurance Institute for Highway Safety whose input improved this paper. She is also grateful to Lisa Henke of the Highway Loss Data Institute, who provided exposure and vehicle feature data, and Jason Rubinoff and JoAnn Wells of the Insurance Institute for Highway Safety for their assistance in obtaining and formatting state crash data, and Raymond Kiefer and Daniel Glaser from General Motors for their efforts in providing GM VINs and the features associated with them. Pennsylvania data used herein were supplied by the Pennsylvania Department of Transportation. The Pennsylvania Department of Transportation specifically disclaims responsibility for any analyses, interpretations, or conclusions drawn in this publication. This work was supported by the Insurance Institute for Highway Safety. 17

19 REFERENCES Bliss, J.P., Gilson, R.D., & Deaton, J.E. (1995). Human probability matching behaviour in response to alarms of varying reliability. Ergonomics, 38(11), doi: / Cicchino, J.B. (2017a). Effects of rearview cameras and rear parking sensors on police-reported backing crashes. Traffic Injury Prevention, 18(8), doi: / Cicchino, J.B. (2017b). Effectiveness of forward collision warning and autonomous emergency braking systems in reducing front-to-rear crash rates. Accident Analysis & Prevention, 99(Pt A), doi: /j.aap Cicchino, J.B. (2018). Real-world effects of General Motors Rear Automatic Braking, Rear Vision Camera, and Rear Parking Assist systems. Arlington, VA: Insurance Institute for Highway Safety. Cicchino, J.B., & McCartt, A.T. (2015). Experiences of model year 2011 Dodge and Jeep owners with collision avoidance and related technologies. Traffic Injury Prevention, 16(3), doi: / Fildes, B., Keall, M., Bos, N., Lie, A., Page, Y., Pastor, C., Pennisi, L., Rizzi, M., Thomas, P., & Tingvall, C. (2015). Effectiveness of low speed autonomous emergency braking in real-world rear-end crashes. Accident Analysis & Prevention, 81, doi: /j.aap Flannagan, C.A., Kiefer, R.J., Bao, S., Leblanc, D.J., & Geisler, S.P. (2014). Reduction of backing crashes by production rear vision camera systems. Journal of Ergonomics, S3, 1 4. doi: / s3-008 Highway Loss Data Institute. (2004). Insurance special report: Standard versus nonstandard risk insurance losses. Arlington, VA. Highway Loss Data Institute. (2014). Evaluation of motorcycle antilock braking systems. HLDI Bulletin, 31(11). Highway Loss Data Institute. (2017). Mazda collision avoidance features: HLDI Bulletin, 34(8). Highway Loss Data Institute. (2018). Unpublished data. Arlington, VA. Keall, M.D., Fildes, B., & Newstead, S. (2017). Real-world evaluation of the effectiveness of reversing camera and parking sensor technologies in preventing backover pedestrian injuries. Accident Analysis & Prevention, 99(Pt A), doi: /j.aap Kidd, D.G., & Brethwaite, A. (2014). Visibility of children behind model year passenger vehicles using glances, mirrors, and backup cameras and parking sensors. Accident Analysis & Prevention, 66, doi: /j.aap Kidd, D.G., Hagoski, B.K., Tucker, T.G., & Chiang, D.P. (2015). The effectiveness of a rearview camera and parking sensor system alone and combined for preventing a collision with an unexpected stationary or moving object. Human Factors, 57(4), doi: /

20 Llaneras, R.E., Green, C.A., Kiefer, R.J., Chundrlik Jr, W.J., Altan, O.D., & Singer, J.P. (2005). Design and evaluation of a prototype rear obstacle detection and driver warning system. Human Factors, 47(1), doi: / Mazzae, E.N., & Barickman, F. (2009). Direct rear visibility of passenger cars: Laser-based measurement development and findings for late model vehicles (Publication No. DOT HS ).Washington, DC: National Highway Traffic Safety Administration. National Highway Traffic Safety Administration. (2016). National Automotive Sampling System/General Estimates System [Data set]. Retrieved from ftp://ftp.nhtsa.dot.gov/ges/ Neurauter, M.L., Llaneras, R.E., Li, B., & Green, C. (2011). Issues related to the use and design of a backing rear cross traffic alert system. SAE International Journal of Passenger Cars-Mechanical Systems, 4(1), doi: / Perez, M., Angell, L.S., Hankey, J., Deering, R.K., Llaneras, R.E., Green, C.A., Neurauter, M.L., & Antin, J.F. (2011). Advanced Crash Avoidance Technologies (ACAT) program final report of the GM-VTTI backing crash countermeasures project (Publication No. DOT HS ).Washington, DC: National Highway Traffic Safety Administration. Reagan, I.J., Cicchino, J.B., Kerfoot, L.B., & Weast, R.A. (2018). Crash avoidance and driver assistance technologies are they used? Transportation Research Part F: Traffic Psychology and Behaviour, 52, doi: /j.trf Reimer, B., Mehler, B., & Coughlin, J.F. (2010). An evaluation of driver reactions to new vehicle parking assist technologies developed to reduce driver stress. Cambridge, MA: MIT Age Lab. 19