inadaassociationbetweenautomaticemergency2025.md 54 KB
category: literaturenote citekey: inadaassociationbetweenautomaticemergency2025 title: Association between automatic emergency braking and pedestrian and cyclist injury severity in Japan authors: "Inada, Haruhiko; Ichikawa, Masao" year: 2025 date: 2025-00-00 2025 doi: 10.1016/j.aap.2025.108091 publication: "Accident Analysis & Prevention" url: "https://doi.org/10.1016/j.aap.2025.108091" zotero_key: MRMCWPRS zotero_storage: 37JAFNEU collections: transpordi_planeerimine folder: 001_artiklid firstAuthor: "Inada, Haruhiko"
status: converted
Contents lists available at ScienceDirect
Accident Analysis and Prevention
journal homepage: www.elsevier.com/locate/aap
Association between automatic emergency braking and pedestrian and cyclist injury severity in Japan
Haruhiko Inada a,* , Masao Ichikawa b
- a Department of Public Health and Health Policy, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan
- b Institute of Medicine, University of Tsukuba, Tsukuba, Ibaraki, Japan
ARTICLE INFO
Keywords: Cross-sectional study Motor vehicle crash Automatic emergency braking Pedestrian Cyclist Injury severity
ABSTRACT
This cross-sectional study assessed whether automatic emergency braking (AEB) in colliding passenger cars is associated with reduced injury severity among pedestrians and cyclists in road traffic crashes. We used national police-reported data in Japan from 2016 to 2019 on vehicle versus pedestrian and vehicle versus cyclist injury crashes, focusing on six best-selling car models with optional AEB systems. We analyzed the association between AEB presence and injury severity, dichotomized as fatal/serious versus minor/no injury, using multivariable logistic regression, adjusting for potential confounders. Among 4131 pedestrians and 6659 cyclists involved in crashes, AEB was associated with a 20 % decrease in the odds of fatal or serious injuries among pedestrians (adjusted odds ratio: 0.80, 95 % confidence interval [CI]: 0.64–0.996). No significant association was observed for cyclists (adjusted odds ratio: 0.91, 95 % CI: 0.74–1.14). In conclusion, AEB systems may mitigate injury severity even when collisions cannot be avoided, at least for pedestrians in real-world conditions. Future studies should evaluate the effectiveness of newer AEB systems that detect cyclists and explore variations in effectiveness according to driver characteristics.
1. Introduction
Globally, the burden of road traffic injuries remains unacceptably high with an estimated deaths of 1.19 million annually or 15 deaths per 100,000 population (World Health Organization, 2023). Pedestrians and cyclists are vulnerable road users and account for 29 % of these deaths (World Health Organization, 2023). Although Japan has a lower road mortality rate at 2.7 deaths per 100,000 population, pedestrians and cyclists comprise approximately half of recent road fatalities; for example, among 3263 road deaths within 30 days of a crash in 2023, 1211 (37 %) were pedestrians 500 (15 %) were cyclists (National Police Agency, 2024).
Vehicle safety technologies offer a promising solution to mitigate the burden, and an increasing number of motor vehicles are equipped with collision avoidance systems, such as automatic emergency braking (AEB). A substantial proportion of pedestrian deaths result from frontal collisions, and many AEB systems are designed to detect pedestrians in front of a vehicle. In Japan, all new passenger cars with up to 10 seats and new freight vehicles weighing *<*3.5 tons are gradually required to have an AEB system since November 2021 (Ministry of Land, Infrastructure, Transport and Tourism of Japan, 2020). The European Union, Australia, and the United States have similar regulations, starting in July 2022, March 2023, and September 2029, respectively (ANCAP, 2021; European Commission, n.d.; NHTSA, 2024).
Studies have demonstrated that AEB can prevent motor vehicle crashes under real-world conditions (Cicchino, 2017, 2022, 2023; Fildes et al., 2015; Isaksson-Hellman & Lindman, 2016; Kinoshita, 2018, 2019; Kondo, 2019; Kullgren et al., 2023; Leslie et al., 2021; Ohlin et al., 2017; Spicer et al., 2018; Teoh, 2021; Yoshida, 2022). Several simulationbased studies have also examined whether AEB in passenger cars reduces injury severity among pedestrians or cyclists (Edwards et al., 2014; Ros´en et al., 2010; Searson et al., 2014). However, to our knowledge, there is only one published study that has examined the association in real-world conditions while controlling for multiple potential confounders, and it did not find statistical evidence for the association (Cicchino, 2022). Its sample size was 649, and the wide confidence interval for the effect estimate (adjusted odds ratio: 1.09; 95 % confidence interval [CI]: 0.59, 2.00) hindered making definitive conclusions. We believe it is valuable to provide epidemiological evidence using real-world data because simulation results may not always reflect real-world outcomes. If AEB reduces injury severity in real-world conditions, its effectiveness in reducing road injuries may be
E-mail addresses: hinada@m.u-tokyo.ac.jp (H. Inada), masao@md.tsukuba.ac.jp (M. Ichikawa).
https://doi.org/10.1016/j.aap.2025.108091
Received 16 September 2024; Received in revised form 28 April 2025; Accepted 30 April 2025 Available online 10 May 2025
* Corresponding author.
underestimated if only its crash avoidance effects are considered. Therefore, this study aimed to examine whether AEB presence in colliding passenger cars is associated with injury severity among collided pedestrians and cyclists. We hypothesized that AEB reduces injury severity by lowering crash speed, even in cases where the crash is not avoided.
2. Materials and methods
2.1. Study design
This was a cross-sectional study. Using police-reported data of vehicle versus pedestrian and vehicle versus bicycle injury crashes in Japan from 2016 to 2019, we examined the association between AEB in colliding passenger cars and injury severity among pedestrians and bicyclists involved in collisions. Throughout the study period, all the AEB systems in this study were expected to detect pedestrians during the daytime, but may fail under certain conditions, such as curved or bumpy roads, tunnels, sudden cut-ins, inclement weather, extremely low or high speed, pedestrians *<*100 cm or > 200 cm in height, groups of multiple pedestrians, and nighttime. By 2018, some AEB systems began to detect cyclists as well; therefore, we analyzed cyclist data in addition to pedestrian data.
2.2. Study sample
We obtained police-reported data on the number of pedestrians and cyclists involved in road traffic crashes via the Institute for Traffic Accident Research and Data Analysis. When a crash is reported to the local police, they collect data according to a standardized protocol, and the institute compiles such data. The data are available upon request after being aggregated by variables, such as year, age and sex of the parties involved, and their injury severity.
In the current study, the data are restricted to vehicle versus pedestrian and vehicle versus cyclist injury crashes from 2016 to 2019 involving six best-selling passenger car models that had optional AEB systems (i.e., there were vehicles with and without AEB in these models) and their drivers were primarily at fault. The six models were ranked among top 20 passenger cars in terms of new vehicle sales in Japan during the study period, and each model had ≥ 99 pedestrian-involved crashes for both AEB-equipped and non-AEB variants during the same period. None of them are sport utility vehicles or pickup trucks. The models were anonymized in the data, and specific models were unidentifiable. The data used in the current study did not include injuries among motor vehicle occupants due to the low incidence of serious injuries or fatalities among them.
2.3. Measures
All variables were categorical with no missing values. The dependent variable was injury severity of pedestrians or cyclists, categorized as fatal, serious, minor, or no injury. Fatalities were recorded at 24 h postcrash. Non-fatal injuries were classified as serious if estimated by a physician to require medical care for ≥ 30 days, and minor if otherwise. The category of no injury exists because it is possible that a motor vehicle occupant was injured, thereby classifying it as an injury crash, while the pedestrian or cyclist was not injured. We dichotomized the variable as fatal or serious versus minor or no injury because the number of fatalities (62 pedestrian and 19 cyclists) and uninjured individuals (0 pedestrians and 17 cyclists) was too small to support multivariable analyses with separate severity levels. The independent variable was AEB in a colliding vehicle, regardless of whether the system was activated, as this information was unavailable.
The data were also stratified according to transport mode (walking or cycling), sex, and age (0–14, 15–29, 30–49, 50–64, 65–74, and 75+ years) of the pedestrian or cyclist; anonymized car model (A–F), sex, age (18–24, 25–64, 65–69, 70–74, 75–79, and 80+ years), and speed at evasive maneuver (0–30 and 31+ km/h) of the driver at fault; and time of the day (day and night), weather (sunny/cloudy and rain/snow/fog), and road surface conditions (dry and wet/frozen/snow) at the time of crash. We consider these variables as potential confounders for the association between the independent and dependent variables. Speed at evasive maneuver (kiken ninchi sokudo in Japanese) is defined as vehicle speed when the driver at fault perceived a potential danger and initiated any evasive maneuvers, such as braking and changing lanes, to avoid the crash. If the driver did not perceive any danger until the crash occurred, it is defined as the vehicle speed right before the crash. An attending police officer estimates the speed at evasive maneuver based on statements from involved parties and eyewitnesses, as well as on-site skid marks.
2.4. Statistical analysis
We conducted all analyses separately for pedestrians and cyclists. After describing the data, including unadjusted logistic regression analyses between the dependent variable and either the independent variable or each covariate, we conducted multivariable logistic regression to examine the association between AEB in the colliding vehicle and injury severity among pedestrians and bicyclists, controlling for potential confounders.
We considered multiplicative two-way interactions between the independent variable and six covariates: the car model, age of the driver at fault, speed at evasive maneuver of the driver at fault, time of the day, weather, and road surface conditions. It is possible that the effectiveness of AEB is different among different car models. A previous study suggested that the effectiveness of AEB may differ among drivers of different age groups (Kondo, 2019). Some manufacturers claim that their AEB may not function fully at low or high speed, at night, or in poor weather conditions. We built a multivariable logistic regression model without interaction terms and separate models each including one of the six covariates and selected the best model based on the Akaike Information Criterion (AIC). Additionally, we repeated the analyses without speed at evasive maneuver because it may be a mediator rather than a confounder between AEB and injury severity, as drivers might perceive potential danger after AEB activation.
We evaluated the goodness-of-fit using the Hosmer–Lemeshow test. We also drew receiver-operating characteristic (ROC) curves and estimated area under the curve (AUC) using five-fold cross-validation with 1000 iterations. All statistical analyses were conducted using R ver. 4.1.3.
3. Results
3.1. Descriptive statistics
During the study period, 4131 pedestrians and 6659 cyclists were involved in injury crashes with one of the six car models. Table 1 shows the number and proportion of pedestrians and cyclists stratified by each variable. Of the 4131 pedestrians, 2760 and 1371 were collided by vehicles with and without AEB, respectively. The proportion of pedestrians killed or seriously injured was 16.7 % (461/2760) when the colliding vehicle had AEB compared with 21.3 % (292/1371) without AEB. For cyclists, these proportions were 8.0 % (350/4392) and 8.1 % (184/ 2267), when the colliding vehicle had AEB and without AEB, respectively. Slightly more than half of the pedestrians (2098/4131) and cyclists (3586/6659) were males. Over a third (1457/4131) of the pedestrians were aged ≥ 65 years, while the cyclists were predominantly young and middle-aged adults. Drivers at fault tended to be males, especially when the colliding vehicle lacked AEB. Nearly 10 % (386/ 4131 for pedestrian crashes and 530/6659 for cyclist crashes) of the drivers were aged ≥ 75 years.
Table 1 Characteristics of pedestrians and cyclists, drivers at fault, and environmental conditions with and without AEB in the colliding motor vehicle, n (%).
| Variable | Pedestrians (N = 4131) | Cyclists (N = 6659) | ||
|---|---|---|---|---|
| AEB (n = 2760) |
Non-AEB (n = 1371) |
AEB (n = 4392) |
Non-AEB (n = 2267) |
|
| Injury severity of the | ||||
| pedestrian/cyclist | ||||
| Fatal or serious injury | 461 (16.7) |
292 (21.3) | 350 (8.0) | 184 (8.1) |
| Minor or no injury | 2299 (83.3) |
1079 (78.7) | 4042 (92.0) |
2083 (91.9) |
| Sex of the pedestrian/ | ||||
| cyclist | ||||
| Female | 1354 (49.1) |
679 (49.5) | 2009 (45.7) |
1064 (46.9) |
| Male | 1406 (50.9) |
692 (50.5) | 2383 (54.3) |
1203 (53.1) |
| Age of the pedestrian/ | ||||
| cyclist | ||||
| 0–14 | 373 (13.5) |
191 (13.9) | 475 (10.8) |
203 (9.0) |
| 15–29 | 352 (12.8) |
170 (12.4) | 1505 (34.3) |
743 (32.8) |
| 30–49 | 571 (20.7) |
278 (20.3) | 1041 (23.7) |
545 (24.0) |
| 50–64 | 499 | 240 (17.5) | 579 | 289 (12.7) |
| 65–74 | (18.1) 453 |
195 (14.2) | (13.2) 412 (9.4) |
254 (11.2) |
| 75+ | (16.4) 512 |
297 (21.7) | 380 (8.7) | 233 (10.3) |
| Car model of the driver at fault |
(18.6) | |||
| A | 1077 | 254 (18.5) | 1654 | 454 (20.0) |
| B | (39.0) 350 |
99 (7.2) | (37.7) 548 |
174 (7.7) |
| C | (12.7) 240 (8.7) |
365 (26.6) | (12.5) 407 (9.3) |
622 (27.4) |
| D | 679 (24.6) |
136 (9.9) | 1168 (26.6) |
220 (9.7) |
| E F |
220 (8.0) 194 (7.0) |
101 (7.4) 416 (30.3) |
345 (7.9) 270 (6.1) |
181 (8.0) 616 (27.2) |
| Sex of the driver at fault | ||||
| Female | 1343 (48.7) |
556 (40.6) | 2166 (49.3) |
979 (43.2) |
| Male | 1417 (51.3) |
815 (59.4) | 2226 (50.7) |
1288 (56.8) |
| Age of the driver at fault | ||||
| 18–24 | 192 (7.0) | 94 (6.9) | 339 (7.7) | 156 (6.9) |
| 25–64 | 1733 (62.8) |
919 (67.0) | 2871 (65.4) |
1590 (70.1) |
| 65–69 | 309 | 138 (10.1) | 438 | 202 (8.9) |
| 70–74 | (11.2) 252 (9.1) |
108 (7.9) | (10.0) 366 (8.3) |
167 (7.4) |
| 75–79 | 169 (6.1) | 63 (4.6) | 256 (5.8) | 97 (4.3) |
| 80+ | 105 (3.8) | 49 (3.6) | 122 (2.8) | 55 (2.4) |
| Speed at evasive maneuver of the driver |
||||
| at fault | ||||
| 0 to 30 km/h | 2429 (88.0) |
1181 (86.1) | 4088 (93.1) |
2112 (93.2) |
| 31+ km/h | 331 (12.0) |
190 (13.9) | 304 (6.9) | 155 (6.8) |
| Time of the day | ||||
| Day | 1727 (62.6) |
858 (62.6) | 3456 (78.7) |
1806 (79.7) |
| Night | 1033 (37.4) |
513 (37.4) | 936 (21.3) |
461 (20.3) |
| Weather | ||||
| Sunny/cloudy | 2349 | 1127 (82.2) | 1991 | 2064 (91.0) |
| Rain/snow/fog | (85.1) 411 |
244 (17.8) | (90.9) 401 (9.1) |
203 (9.0) |
| (14.9) | ||||
| Road surface conditions Dry |
2287 (82.9) |
1075 (78.4) | 1923 (89.3) |
2036 (89.8) |
Table 1 (continued )
| Pedestrians (N = 4131) | Cyclists (N = 6659) | |||
|---|---|---|---|---|
| Variable | AEB (n = 2760) |
Non-AEB (n = 1371) |
AEB (n = 4392) |
Non-AEB (n = 2267) |
| Wet/frozen/snow | 473 (17.1) |
296 (21.6) | 469 (10.7) |
231 (10.2) |
AEB: automatic emergency braking.
3.2. Logistic regression
Tables 2 and 3 present the results of unadjusted and multivariable (i. e., adjusted) logistic regression analyses for the pedestrians and cyclists, respectively. For pedestrians, the model with the interaction term between AEB and speed at evasive maneuver was selected, and for cyclists, the model with the interaction term between AEB and weather was selected. Odds ratio for the association between AEB and death or serious injury among pedestrians was 0.74 (95 % CI: 0.63, 0.87) in the unadjusted analysis and 0.80 (95 % CI: 0.64, 0.996) in the multivariable analysis, indicating a 20 % reduction in the odds of death or serious injury among pedestrians when the colliding vehicle had AEB compared with when it did not. For cyclists, the odds ratios were 0.98 (95 % CI: 0.81, 1.18) in the unadjusted analysis and 0.91 (95 % CI: 0.74, 1.14) in the multivariable analysis. In the supplementary material, Tables S1 and S2 provide raw regression coefficients before exponentiation and their standard errors, and Figs. S1 and S2 show the cross-validated ROC curves for the pedestrian and cyclist models. Tables S3 and S4 present the results of the multivariable logistic regression analyses for all interactions except those with the lowest AIC, which are shown in Tables 2 and 3. All models yielded odds ratios for the association between AEB and death or serious injury that were similar to those in the main models for pedestrians and cyclists, respectively. Excluding the speed at evasive maneuver from the main models changed the results minimally (Tables S5 and S6).
4. Discussion
4.1. Principal findings
This cross-sectional study found that AEB presence in colliding passenger cars was associated with lower injury severity among collided pedestrians but not cyclists after controlling for multiple potential confounders. To our knowledge, this is the first study to demonstrate an inverse association between AEB in colliding passenger cars and injury severity among collided pedestrians using real-world data.
4.2. Strengths and weaknesses
The current study has several strengths. Our study population is a large sample from the national police-reported data of road traffic injuries of pedestrians and cyclists. They were collided by six models of passenger cars with optional AEB during the study period, which enabled us to compare injury severity of those collided by vehicles of the same model with and without AEB. We controlled for multiple potential confounders, including the sex and age of the pedestrians and cyclists and speed at evasive maneuver of the drivers. Consistency of the point estimates between the unadjusted and adjusted models for both pedestrians and cyclists suggest minimal confounding. Our findings align with previous studies that showed AEB prevents motor vehicle crashes (Cicchino, 2017, 2022, 2023; Fildes et al., 2015; Isaksson-Hellman & Lindman, 2016; Kinoshita, 2018, 2019; Kondo, 2019; Kullgren et al., 2023; Leslie et al., 2021; Ohlin et al., 2017; Spicer et al., 2018; Teoh, 2021; Yoshida, 2022). Our sample size for pedestrians (4131) was substantially larger than that of a previous study that examined the association between AEB and injury severity (Cicchino, 2022), which led to more precise estimates. It is noteworthy that there was no statistical
Table 2 Association between AEB in the colliding motor vehicle and death or serious injury of pedestrians (n = 4131).
| Point estimate of the odds ratio (95 % confidence interval) |
||
|---|---|---|
| Unadjusted | Adjusted (AUC: 0.74) |
|
| AEB | ||
| Yes | 0.74 (0.63, | 0.80 (0.64, |
| 0.87) | 0.996) | |
| No | Reference | Reference |
| Sex of the pedestrian | ||
| Female Male |
Reference 0.68 (0.58, 0.80) |
Reference 0.80 (0.67, 0.95) |
| Age of the pedestrian | ||
| 0–14 | 1.51 (1.08, 2.13) |
1.44 (1.00, 2.08) |
| 15–29 | 0.75 (0.50, 1.14) |
0.66 (0.43, 1.01) |
| 30–49 | Reference | Reference |
| 50–64 | 1.91 (1.41, 2.60) |
2.06 (1.50, 2.84) |
| 65–74 | 3.68 (2.74, 4.93) |
3.74 (2.74, 5.09) |
| 75+ | 5.35 (4.06, 7.06) |
5.28 (3.93, 7.09) |
| Car model of the driver at fault | ||
| A B |
Reference 1.21 (0.94, |
Reference 1.25 (0.94, 1.66) |
| C | 1.57) 1.17 (0.92, 1.48) |
0.97 (0.74, 1.28) |
| D | 0.67 (0.53, 0.86) |
0.81 (0.62, 1.06) |
| E | 0.75 (0.54, 1.05) |
0.86 (0.60, 1.24) |
| F | 0.83 (0.64, 1.07) |
0.81 (0.60, 1.09) |
| Sex of the driver at fault | ||
| Female | Reference | Reference |
| Male | 0.99 (0.85, 1.17) |
1.17 (0.97, 1.40) |
| Age of the driver at fault | ||
| 18–24 | 1.19 (0.88, | 1.13 (0.81, 1.59) |
| 1.61) | ||
| 25–64 65–69 |
Reference 0.89 (0.68, |
Reference 0.80 (0.59, 1.08) |
| 1.17) | ||
| 70–74 | 1.23 (0.93, 1.61) |
1.17 (0.86, 1.57) |
| 75–79 | 1.04 (0.74, 1.47) |
0.91 (0.62, 1.32) |
| 80+ | 0.97 (0.64, 1.49) |
0.88 (0.56, 1.39) |
| Speed at evasive maneuver of the driver at fault |
||
| 0 to 30 km/h | Reference | Reference |
| 31+ km/h | 5.37 (4.41, 6.53) |
6.66 (4.67, 9.50) |
| Time of the day | ||
| Day Night |
Reference 1.42 (1.21, 1.66) |
Reference 1.29 (1.07, 1.56) |
| Weather | ||
| Sunny/cloudy | Reference | Reference |
| Rain/snow/fog | 0.88 (0.70, 1.10) |
0.98 (0.64, 1.50) |
| Road surface | ||
| Dry | Reference | Reference |
| Wet/frozen/snow | 0.90 (0.74, 1.11) |
0.98 (0.66, 1.46) |
| Interaction | ||
| AEB*Speed 31+ km/h AIC score |
NA NA |
0.71 (0.46, 1.10) 3392.4 |
| Hosmer–Lemeshow test | NA | 0.41 |
AEB: automatic emergency braking; AIC: Akaike Information Criterion; AUC: Area under the curve; NA: Not applicable.
evidence that AEB was associated with lower injury severity among cyclists. This is not only because it is consistent with the fact that AEB under investigation in this study did not detect cyclists until around 2018 but also it indicates that a potential confounding of drivers of passenger cars with an AEB system are safer than those of cars without an AEB system is small; this means that the observed association among pedestrian injury severity is unlikely due to this potential confounding. We consider that the generalizability of our findings depends on AEB systems, road traffic situations, and vehicle regulations in countries in particular, but since the same AEB systems are expected to function similarly, AEB systems of at least the six passenger car models that were examined in this study would have similar effects in many countries.
Our study also has limitations. The observational nature of the study limits our ability to infer causality, though the temporality of AEB and injury severity cannot be reversed. Additionally, as explained in the Methods section, speed at evasive maneuver may be a mediator rather than a confounder. However, the change in the estimated association in the multivariable model with the lowest AIC was minimal after excluding speed at evasive maneuver from the model; therefore, we consider that speed at evasive maneuver is significant neither as a confounder nor a mediator and that this weakness does not materially affect our findings. It is also possible that motor vehicle crashes that occur despite activated AEB tend to be high-speed incidents that AEB could not prevent. If so, pedestrians and cyclists collided by vehicles with AEB may sustain more serious injuries than those collided by vehicles without AEB. This results in a bias in the association between AEB and injury severity upwards, i.e., associating AEB with more serious injury; consequently, the unbiased association is likely smaller than our estimates. We also acknowledge that some AEB systems may have been updated and improved during the study period, but we were unable to account for this in our analysis. However, this limitation does not undermine our findings because newer AEBs are presumably more effective in reducing crash severity than older ones; therefore, our results likely underestimate their effectiveness.
4.3. Meaning of the study
Our study results support the effectiveness of the six car models' AEB for detecting at least pedestrians during the daytime. These systems not only prevent crashes but also mitigates crash severity when collisions occur in real-world conditions. Expanding the use of vehicles equipped with AEB could contribute to reducing the burden of road traffic injuries among pedestrians.
4.4. Unanswered questions and future research
We need to examine whether the latest AEB systems, which are designed to detect cyclists as well, can also mitigate their injuries. Additionally, the magnitude of association between AEB and injury severity may vary across driver age groups. If AEB systems reduce injury severity for collided vulnerable road users even among older drivers, they could be a viable option for extending older drivers' safe driving years and supporting safer mobility in aging societies. This has potential implications for licensing policies and vehicle safety standards. Future research should also explore contextual factors, such as urban versus rural settings, that may influence AEB performance across different driver groups.
5. Conclusions
We found that AEB presence in colliding passenger cars was associated with lower injury severity among collided pedestrians. These findings underscore the potential of this technology to mitigate crash
Table 3 Association between AEB in the colliding motor vehicle and death or serious injury of cyclists (n = 6659).
| | Point estimate of the odds ratio (95
% confidence interval) | | |
|--------------------------------------------|----------------------------------------------------------------|-------------------------|--|
| | Unadjusted | Adjusted
(AUC: 0.70) | |
| AEB | | | |
| Yes | 0.98 (0.81, | 0.91 (0.74, | |
| | 1.18) | 1.14) | |
| No | Reference | Reference | |
| Sex of the cyclist | | | |
| Female | Reference | Reference | |
| Male | 0.90 (0.76, | 0.92 (0.76, | |
| | 1.08) | 1.11) | |
| Age of the cyclist | | | |
| 0–14 | 1.63 (1.15, | 1.37 (0.95, | |
| | 2.30) | 1.96) | |
| 15–29 | 0.82 (0.61, | 0.75 (0.56, | |
| | 1.10) | 1.02) | |
| 30–49 | Reference | Reference | |
| 50–64 | 1.92 (1.40, | 1.91 (1.39, | |
| | 2.62) | 2.63) | |
| 65–74 | 2.66 (1.94, | 2.45 (1.78, | |
| | 3.63) | 3.39) | |
| 75+ | 4.03 (2.99, | 3.54 (2.60, | |
| | 5.42) | 4.84) | |
| Car model of the driver at fault | | | |
| A | Reference | Reference | |
| B | 0.83 (0.61, | 0.75 (0.55, | |
| | 1.12) | 1.04) | |
| C | 0.75 (0.57, | 0.68 (0.50, | |
| | 0.98) | 0.92) | |
| D | 0.60 (0.46, | 0.69 (0.52, | |
| | 0.79) | 0.92) | |
| E | 0.87 (0.62, | 0.95 (0.67, | |
| | 1.22) | 1.36) | |
| F | 0.85 (0.64, | 0.84 (0.61, | |
| | 1.13) | 1.14) | |
| Sex of the driver at fault | | | |
| Female | Reference | Reference | |
| Male | 0.84 (0.71, | 0.90 (0.74, | |
| | 1.01) | 1.10) | |
| Age of the driver at fault | | | |
| 18–24 | 1.38 (1.01, | 1.28 (0.92, | |
| | 1.89) | 1.79) | |
| 25–64 | Reference | Reference | |
| 65–69 | 0.91 (0.64, | 0.94 (0.67, | |
| | 1.26) | 1.32) | |
| 70–74 | 1.13 (0.82, | 1.09 (0.77, | |
| | 1.57) | 1.52) | |
| 75–79 | 1.31 (0.90, | 1.31 (0.89, | |
| | 1.90) | 1.93) | |
| 80+ | 2.02 (1.30, | 2.14 (1.35, | |
| | 3.13) | 3.41) | |
| Speed at evasive maneuver of the driver at | | | |
| fault | | | |
| 0 to 30 km/h | Reference | Reference | |
| 31+ km/h | 5.22 (4.16, | 5.03 (3.94, | |
| | 6.57) | 6.41) | |
| Time of the day | | | |
| Day | Reference | Reference | |
| Night | 0.94 (0.75, | 1.02 (0.80, | |
| | 1.17) | 1.30) | |
| Weather | | | |
| Sunny/cloudy | Reference | Reference | |
| Rain/snow/fog | 1.01 (0.75, | 0.71 (0.33, | |
| | 1.38) | 1.52) | |
| Road surface | | | |
| Dry | Reference | Reference | |
| Wet/ frozen/snow | 1.06 (0.80, | 1.10 (0.65, | |
| | 1.41) | 1.85) | |
| Interaction | | | |
| AEB*Rain/snow/fog weather | NA | 1.75 (0.86, | |
| | | 3.56) | |
| AIC score | NA | 3430.6 | |
| | | | |
AEB: automatic emergency braking; AIC: Akaike Information Criterion; AUC: Area under the curve; NA: Not applicable.
severity in real-world conditions.
Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work the authors used ChatGPT-4o in order to improve readability and language of the manuscript. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.
CRediT authorship contribution statement
Haruhiko Inada: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology, Formal analysis, Data curation. Masao Ichikawa: Writing – review & editing, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research was funded by the Japan Society for the Promotion of Science KAKENHI (grant numbers 21H03195 and 23K21516). We thank the Institute for Traffic Accident Research and Data Analysis for providing the data. We would also like to thank Editage (www.editage. com) for English language editing. This study did not require approval from the institutional review board because it is a cross-sectional study that used only aggregate data from the public domain.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.aap.2025.108091.
Data availability
The authors do not have permission to share data.
References
ANCAP. (2021, November 11). ANCAP welcomes mandatory autonomous emergency braking (AEB) technology from 2023. https://www.ancap.com.au/media-and-gallery/ media-releases/dbf7f0.
Cicchino, J.B., 2017. Effectiveness of forward collision warning and autonomous emergency braking systems in reducing front-to-rear crash rates. Acc. Anal. Prevent. 99, 142–152. https://doi.org/10.1016/J.AAP.2016.11.009.
Cicchino, J.B., 2022. Effects of automatic emergency braking systems on pedestrian crash risk. Acc. Anal. Prevent. 172, 106686. https://doi.org/10.1016/J. AAP.2022.106686.
Cicchino, J.B., 2023. Effects of forward collision warning and automatic emergency braking on rear-end crashes involving pickup trucks. Traffic Inj. Prev. 24 (4), 293–298. https://doi.org/10.1080/15389588.2023.2176191.
Edwards, M., Nathanson, A., Wisch, M., 2014. Estimate of potential benefit for Europe of fitting Autonomous Emergency Braking (AEB) systems for pedestrian protection to passenger cars. Traffic Inj. Prev. 15 (Suppl 1), S173–S182. https://doi.org/10.1080/ 15389588.2014.931579.
European Commission. (n.d.). Vehicle Safety and automated/connected vehicles. https:// single-market-economy.ec.europa.eu/sectors/automotive-industry/vehicle-safetyand-automatedconnected-vehicles_en#:~:text=The%20revised%20General% 20Safety%20Regulation,automated%20vehicles%20in%20the%20EU.
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. Acc. Anal. Prevent. 81, 24–29. https://doi.org/ 10.1016/J.AAP.2015.03.029.
- Isaksson-Hellman, I., Lindman, M., 2016. Evaluation of the crash mitigation effect of lowspeed automated emergency braking systems based on insurance claims data. Traffic Inj. Prev. 17 (Suppl 1), 42–47. https://doi.org/10.1080/15389588.2016.1186802.
- Kinoshita, Y. (2018). AEB ni yoru tsuitotsu jiko teigen koka no bunseki (Analysis of the effectiveness of AEB in reducing rear-end collisions). [in Japanese]. https://www. itarda.or.jp/presentation/21/show_lecture_file.pdf?lecture_id=110&type=file_jp.
- Kinoshita, Y. (2019). Shototsu higai keigen brake (AEB) no tsuitotsu jiko teigen koka hosoku bunseki (Supplementary analysis of the effectiveness of AEB in reducing rearend accidents). [in Japanese]. https://www.itarda.or.jp/presentation/22/show_ lecture_file.pdf?lecture_id=124&type=file_jp.
- Kondo, N. (2019). Shotsu higai keigen brake (AEB) no sedai betsu koka bunseki (Analysis of the effectiveness of AEB by its generation). [in Japanese]. https://www.itarda.or. jp/presentation/22/show_lecture_file.pdf?lecture_id=125&type=file_jp.
- Kullgren, A., Amin, K., Tingvall, C., 2023. Effects on crash risk of automatic emergency braking systems for pedestrians and bicyclists. Traffic Inj. Prev. 24 (Suppl 1), S111–S115. https://doi.org/10.1080/15389588.2022.2131403.
- Leslie, A.J., Kiefer, R.J., Meitzner, M.R., Flannagan, C.A., 2021. Field effectiveness of general motors advanced driver assistance and headlighting systems. Acc. Anal. Prevent. 159, 106275. https://doi.org/10.1016/J.AAP.2021.106275.
- Ministry of Land, Infrastructure, Transport and Tourism. (2020, January 31). Adopting international standards for automatic emergency braking systems for passenger cars and making it mandatory for new cars. [in Japanese]. https://www.mlit.go.jp/report/ press/jidosha08_hh_003618.html.
National Police Agency. (2024). Reiwa 5 nenchu no 30 nichi inai kotsu jiko shisha no jokyo. (Road traffic deaths within 30 days of a crash in 2023). [in Japanese]. https://www.estat.go.jp/stat-search/file-download?statInfId=000040162340&fileKind=2.
NHTSA. (2024, April 29). NHTSA Finalizes Key Safety Rule to Reduce Crashes and Save Lives. https://www.nhtsa.gov/press-releases/nhtsa-fmvss-127-automaticemergency-braking-reduce-crashes#:~:text=Today%2C%20the%20U.S.% 20Department%20of,light%20trucks%20by%20September%202029.
Ohlin, M., Strandroth, J., Tingvall, C., 2017. The combined effect of vehicle frontal design, speed reduction, autonomous emergency braking and helmet use in reducing real life bicycle injuries. Saf. Sci. 92, 338–344. https://doi.org/10.1016/J. SSCI.2016.05.007.
Ros´en, E., K¨ allhammer, J.E., Eriksson, D., Nentwich, M., Fredriksson, R., Smith, K., 2010. Pedestrian injury mitigation by autonomous braking. Acc. Anal. Prev. 42 (6),
Searson, D.J., Anderson, R.W.G., Hutchinson, T.P., 2014. Integrated assessment of pedestrian head impact protection in testing secondary safety and autonomous emergency braking. Acc. Anal. Prev. 63, 1–8. https://doi.org/10.1016/J. AAP.2013.10.014.
Spicer, R., Vahabaghaie, A., Bahouth, G., Drees, L., Martinez von Bülow, R., Baur, P., 2018. Field effectiveness evaluation of advanced driver assistance systems. Traffic Inj. Prev. 19 (Suppl 2), S91–S95. https://doi.org/10.1080/ 15389588.2018.1527030.
Teoh, E.R., 2021. Effectiveness of front crash prevention systems in reducing large truck real-world crash rates. Traffic Inj. Prev. 22 (4), 284–289. https://doi.org/10.1080/ 15389588.2021.1893700.
World Health Organization. (2023). Global Status Report on Road Safety 2023. https:// www.who.int/publications/i/item/9789240086517.
Yoshida, S. (2022). Kosoku doro ni okeru tsuitotsu jiko hassei jokyo to AEB no koka bunseki (Analysis of rear-end collision incidents and the effectiveness of AEB on highways). [in Japanese]. https://www.itarda.or.jp/presentation/25/show_lecture_ file.pdf?lecture_id=146&type=file_jp.