SWOV Catalogus

341411

Using naturalistic driving data to assess the prevalence of environmental factors and driver behaviors in teen driver crashes.
20150878 ST [electronic version only]
Carney, C. McGehee, D. Harland, K. Weiss, M. & Raby, M.
Washington, D.C., American Automobile Association AAA Foundation for Traffic Safety, 2015, 79 p., 51 ref.

Samenvatting Objective, detailed and accurate information regarding the prevalence of factors with the potential to contribute to crashes is vital. In the past, the only way to obtain information for a large number of crashes was to use data collected from police reports. While information gathered this way is helpful, it has many limitations. More recently, in-vehicle event recorders (IVERs) have become a widely accepted means of gathering crash data both in research and real-world applications. In this study, we conducted a large-scale comprehensive examination of naturalistic data from crashes that involved teenage drivers. Other naturalistic studies have investigated only a small number of crashes or used near crashes as a proxy for actual crashes, and few crashes involving teen drivers have been observed in other naturalistic studies. In contrast, this project examined naturalistic data from thousands of actual crashes that involved teenage drivers. The data allowed us to examine behaviours and potential contributing factors in the seconds leading up to the collision, and provided information not available in police reports. A coding method was developed specifically for this study, and video data were coded with the goal of identifying the factors present prior to crashes–in particular the prevalence of potentially distracting driver behaviours and drowsiness. The study addressed the following research questions: • What were the roadway and environmental conditions at the time of the crash? • What were the critical events and potential contributing factors leading up to the crash and did these differ by crash type? • What driver behaviours were present in the vehicle prior to the crash and did these differ by crash type? • How did driver reaction times and eyes-off-road time differ relative to certain driver behaviours and crash types? • Could drowsy driving be detected using this type of crash data? Understanding the prevalence of potential contributing causes of crashes provides a significant societal benefit and advances the field of traffic safety. More specifically, information regarding what is happening inside the vehicle during the seconds before a crash can suggest countermeasures such as education, training, or advanced safety technologies that might best mitigate certain types of crashes. Lytx, a company that has been collecting data using in-vehicle event recorders (IVERs) for over a decade, provided the crash data. Their DriveCam system collects video, audio and accelerometer data when a driver triggers the device by hard braking, fast cornering, or an impact that exceeds a certain g-force. Each video is 12-seconds long, and provides information on the 8 seconds before and 4 seconds after the trigger. The system has a wide range of applications–families use them to help young drivers as they begin to drive independently, while over 500 commercial and government fleets employ them for fleet management. Crashes examined in this study involved drivers aged 16-19 who were participating in a teen driving program that involved the use of a DriveCam system. Ltyx made 6,842 videos of crashes that occurred between August 2007 and July 2013 available for review. In order to reduce this number and to eliminate minor curb strikes from the analysis, those crashes in which the vehicle sustained forces less than 1g were excluded. Crashes in which the DriveCam equipped vehicle was struck from behind were excluded. Additional videos were excluded for other reasons (e.g., animal strikes, video problems, or the driver not being a teen). A total of 1,691 moderate-to-severe crashes met the inclusion criteria and were analysed for the current study. Video from the 6 seconds preceding each crash were coded for analysis. A coding methodology which focused on identifying the factors present in crashes was developed specifically for gathering information from the videos. Data elements coded for each crash included environmental conditions, contributing circumstances (e.g., inadequate surveillance, running traffic signals), and driver and passenger behaviours. Each crash was double coded by two University of Iowa (UI) analysts and mediated by a third when necessary. For this study, 1,691 moderate-to-severe crashes involving young drivers ages 16-19 were reviewed. Of these crashes, 727 were vehicle-to-vehicle crashes in which the force of the impact was 1.0g or greater, and 964 were single-vehicle crashes in which the vehicle’s tires left the roadway and impacted (with a force of 1.0g or greater) one or more natural or artificial objects. While the extent of any injuries sustained in the crashes was not evident from the videos, it is known that no fatal crashes were included in this analysis. Additionally, while it is likely that most of the vehicle-to-vehicle crashes in the analyses resulted in a police report being filed, many of the single-vehicle crashes may have gone unreported. Characteristics of drivers and passengers: Male drivers were involved in 52% of crashes and females 48%. When drivers were examined by crash type, results indicated that more males were involved in single-vehicle crashes than females (56% vs 44%), and more females in vehicle-to-vehicle crashes than males (53% vs 46%). The driver was seen wearing a seatbelt in 93% of all crashes. Passengers were present in the vehicle in one-third of crashes (36%), with one passenger present in 25.5% and two or more passengers present in 10.5%. One-quarter (27%) of crashes with passengers showed at least one passenger that was unbelted. The majority of passengers, when present, were estimated to be 16-19 years old (84%); 55% of the passengers were male. Characteristics of roadway and environment: In general, crashes occurred most often on roadways that connect local streets, called collectors (52%). However, when examined by crash type, single-vehicle crashes were more likely to occur on collectors than vehicle-to-vehicle crashes (66% vs 35%), and vehicle-to-vehicle crashes were more likely than single-vehicle crashes to occur on arterials (47% vs 8%). Road surface conditions were more likely to be dry for vehicle-to-vehicle crashes than for single-vehicle crashes (79% vs 19%); a much greater proportion of single-vehicle crashes than vehicle-to-vehicle occurred on roads covered with snow or ice (65% vs 8%). Overall, 60% of crashes occurred when there was no adverse weather; however, this was significantly more likely to be the case for vehicle-to-vehicle crashes than for single-vehicle crashes (74% vs 48%). Vehicle-to-vehicle crashes were more likely to happen during the week than single-vehicle crashes (79% vs 65%), with more occurring on Friday than any other day. In addition, vehicle-to-vehicle crashes were significantly more likely than single-vehicle crashes to occur between 3pm and 6pm (36% vs 19%). In contrast, single-vehicle crashes were more likely to occur on a weekend (35% vs 21%) and nearly three times as likely to occur between 9pm and midnight (14% vs 5%). Characteristics of crashes: Recognition errors (e.g., inattention and inadequate surveillance) and decision errors (e.g., failing to yield right of way, running stop signs and driving too fast) were the most common errors made by young drivers, occurring in 70% and 66% of all crashes, respectively. However, when examined by crash type, recognition errors were significantly more common in vehicle-to-vehicle crashes than in single-vehicle crashes (89% vs 56%). In addition, both performance errors (e.g., losing control and overcorrecting) and decision errors were significantly more frequent in single-vehicle crashes (82% vs 9%, and 80% vs 47%, respectively). Characteristics of vehicle-to-vehicle crashes: The majority of vehicle-to-vehicle crashes were rear-end (57%) and angle (40%) crashes. Eighty-eight percent of rear-end crashes in which the DriveCam-equipped vehicle struck a lead vehicle involved another vehicle in the driver’s lane decelerating or stopping on the roadway. (Rear-end crashes in which the DriveCam-equipped vehicle was struck from behind were not included in this analysis.) Of angle crashes, 58% involved the participant’s vehicle crossing the centerline or turning at an intersection; 38% involved another vehicle encroaching on the participant’s vehicle. Regardless of fault, in 94% of crashes the driver potentially contributed to the crash in some way. Decision errors such as a failure to yield right of way (ROW) and running stop signs/signals were significantly more frequent in angle crashes than in rear-end crashes (61% vs 38%). Recognition errors such as inadequate surveillance and inattention, as well as performance errors such as losing control of the vehicle, were more frequent in rear-end crashes than in angle crashes (93% vs 82%, and 11% vs 5%, respectively). Characteristics of single-vehicle crashes: Of the single-vehicle crashes coded, 66% were loss-of-control (LOC) crashes due to road surface or weather conditions combined with travelling too fast for the conditions; 19% were road-departure crashes attributed to driver inattention due to distraction or inadequate surveillance; 12% were LOC crashes attributed to excessive speed (not related to road or weather conditions); and 3% were LOC due to an evasive manoeuvre. Only one crash was attributed to LOC due to mechanical failure (a brake failure was evident in one crash). Regardless of fault, the driver was considered to have potentially contributed in some way to 99% of the crashes. Recognition errors (i.e., inadequate surveillance or inattention) were present in 100% of road-departure crashes compared to only 46% of LOC crashes. Decision errors such as driving too fast and following too closely were more common in LOC crashes than in road-departure crashes (99% vs 4%). Finally, performance errors such as losing control of the vehicle and overcorrecting/over steering were also more common in LOC crashes, present nearly 100% of the time, compared to only 12% of road-departure crashes. Driver behaviours: Drivers were seen engaging in some type of potentially distracting behaviour leading up to 58% of all crashes examined. The two most frequently seen driver behaviours were attending to passengers (14.9%) and cell phone use (11.9%). Cell phone use was significantly more likely in road-departure crashes than any other type of crash (34% vs 9.2%). Attending to a passenger was slightly less likely to be seen during a road-departure crash than any other crash types (13.3% vs 15.0%). Overall, males and females were equally likely to be engaged in potentially distracting behaviour. However, females were more likely than males to have been using a cell phone (14% vs 10%), engaged in personal grooming (7% vs. 5%), or singing/dancing to music (9% vs 6%) prior to the crash. Additionally, for all types of crashes, drivers were significantly more likely to have been using their cell phone when they were alone in the vehicle than when they had passengers. Drivers were found to have been looking away from the roadway for a significantly longer length of time prior to the crash in road departure crashes than in any other type of crash; mean eyes-off-road times were 4.0s for road departure crashes, 2.5s for rear-end crashes, 0.7s for angle crashes, and 0.5s for LOC crashes. Of all driver behaviours, using electronic devices, attending to a moving object in the vehicle, using a cell phone and reaching for an object resulted in the longest mean eyes-off- road times (3.9s, 3.6s, 3.3s, and 3.3s, respectively). Drivers engaged in cell phone use had mean eyes-off-road times that were twice as long as those drivers who were attending to passengers (3.3s vs 1.5s). Also, when cell phone use was analysed separately, the average eyes-off-road time for drivers who were operating or looking at their phone was 4.1s, compared to 0.9s for drivers who were talking or listening. Reaction time was analysed for rear-end crashes only. Results found that drivers who were using a cell phone had a significantly longer reaction time than drivers not engaged in any behaviours (2.8s vs 2.1s). In contrast, drivers attending to passengers had similar reaction times to drivers not engaged in any behaviours (2.2s vs 2.1s). In addition, in over 50% of rear-end crashes where the driver was engaged in cell-phone us, the driver showed no reaction at all (braking or steering), whereas the driver failed to react at all in only 9.5% of crashes with a driver attending to a passenger. Passenger behaviours: Passengers were present in 36% of the crashes. The majority of passengers present in the crashes examined were estimated to be 16-19 years old (84%), and 55% were male. Overall, the most frequent behaviour that passengers were seen engaging in was conversation with the driver. When single passengers were present, they were engaged in conversation with the driver 36% of the time, and when two or more passengers were present, 39% of the time. When two or more passengers were present, they were significantly more likely to be making loud noises (5% vs 0.2%), moving around in the vehicle (14% vs 6%) and texting/using cell phone (7% vs 3%) than when only a single passenger was present. Drowsy driving: Determining whether or not a driver was drowsy was extremely difficult given the limitations associated with event-triggered naturalistic driving data. Only 15 of the 1,691 crashes reviewed contained conclusive evidence of drowsy driving; however, it is possible that drowsiness was present in cases in which it could not be ascertained with only 6 seconds of pre-crash video. Use of IVERs in naturalistic driving allows researchers a unique view into the vehicle and provides invaluable information regarding the behavioural and environmental factors present before a crash. The data gathered offers a much more detailed context relative to police reports and other crash databases, and allows more micro-level analyses to be conducted. This study examines the roadway and environmental conditions present in different types of crashes. It describes the critical events and contributing factors that led to crashes, and how they varied by crash type. It also provides information regarding the possible effect certain driver behaviours could have on reaction time and eyes-off-road time. Finally, it is the first and largest naturalistic study of moderate-to-severe crashes to examine driver and passenger behaviours for a variety of crash types. As was expected, environmental and roadway conditions varied considerably by crash type, with single-vehicle crashes being most affected by weather and surface conditions. Time of day also played a role, with single-vehicle crashes being more likely to occur at night, while vehicle-to-vehicle crashes were more likely during times of high traffic flow. Recognition errors were more common for vehicle-to-vehicle crashes, while performance errors were more frequent in single-vehicle crashes. While drivers were seen engaging in a wide range of behaviours leading up to a crash, the most common behaviour among young drivers was attending to passengers. When passengers were present, the most common behaviour they engaged in was conversation with the driver. Cell phone use was also seen frequently for all drivers, with operating/looking at the phone (e.g., texting) observed most often. Interestingly, all drivers were significantly more likely to be using a cell phones (for talking or texting) when they were alone in the vehicle. Cell phone use was more common in road departure crashes and contributed to significantly longer reaction times. Potentially distracting behaviours in general, and cell phone use in particular, were much more prevalent in the current study than in official statistics based on police reports. One unexpected result was that reaction times were not significantly longer when drivers were attending to passengers than when they were not. The results of this study can be used to inform the development of education, training, and technology-based interventions aimed at reducing teen drivers’ crash risk. (Author/publisher)
Full-text
Dossier
Suggestie? Neem contact op met de SWOV bibliotheek voor uw opmerkingen
Copyright © SWOV | Juridisch voorbehoud | Contact