Driver's behavioral adaptation to adaptive cruise control (ACC): the case of speed and time headway

Dashcam video footage-based analysis of microsleep-related behaviors in truck collisions attributed to falling asleep at the wheel

OBJECTIVE

With the rapid spread of dashcams, many car accidents have been recorded; however, behavioral approaches using these dashcam video footage have not been sufficiently examined. We employed dashcam video footage to evaluate microsleep-related behaviors immediately prior to real-world truck collisions in professional drivers to explore a new solution to reduce collisions attributed to falling asleep at the wheel.

METHODS

In total, 3,120 s of video footage (60 s/case × 52 cases) from real-world truck collisions of 52 professional drivers obtained from interior and exterior dashcams were used and visually analyzed in a second-by-second manner to simultaneously evaluate any eye changes and microsleep-related behaviors (the driver's anti-sleepiness behavior, behavioral signs of microsleep, and abnormal vehicle behavior) during driving.

RESULTS

Assessment of the frequency of occurrence of each item of microsleep-related behavior in the 52 collisions revealed that the item "touching" in terms of anti-sleepiness behavior, "absence of body movement" in terms of behavioral signs of microsleep, and "inappropriate line crossing" in terms of abnormal vehicle behavior were observed at the highest rate in all drivers (46.2%, 75.0%, and 78.8%, respectively). Decreases in anti-sleepiness behavior coincided with increases in behavioral signs of microsleep and abnormal vehicle behavior, with collisions occurring within approximately 40 s of these changes. Collisions were more common among young people and in the early morning and evening.

CONCLUSION

Our dashcam video footage-based analysis in truck collisions attributed to falling asleep at the wheel revealed the process of changes in microsleep-related driver and vehicle behaviors, classified as anti-sleepiness behavior, behavioral signs of microsleep, and abnormal vehicle behavior. Based on these findings, to prevent collisions caused by falling asleep at the wheel, it is crucial to monitor not only the driver's eyes, but also the driver's whole body and vehicle behavior simultaneously to reliably detect microsleep-related behaviors.

Sleep-deprived car-following: Indicators of rear-end crash potential

Safety assessment among sleep-deprived drivers is a challenging research area with only a few sleep-related studies investigating safety performance during car-following. Therefore, this study aimed to measure the effects of partial sleep deprivation on driver safety during car-following. Fifty healthy male drivers with no prior history of any sleep-related disorders, drove the driving simulator in three conditions of varying sleep duration: a baseline (no sleep deprivation), test session (TS1) after one night of PSD (sleep ≤4.5 h/night) and TS2 after two consecutive nights of PSD. The reduced sleep in PSD sessions was monitored using an Actiwatch. Karolinska Sleepiness Scale was used to indicate loss of alertness among drivers. Each drive included a car-following task to measure longitudinal safety indicators based on speed and headway management: normalized time exposed to critical gap (TECG'), safety critical time headway and speed variability with respect to leading vehicle's speed (SPV). Crash potential index (CPI) was also determined from deceleration rate of drivers during car-following and was found correlated with other indicators. Therefore, to determine the aggregate influence of PSD on safety during car-following, CPI was modelled in terms of TECG, SPV, THW and other covariates. All safety metrics were modelled using generalized mixed effects regression models. The results showed that compared to the baseline drive, critical time headway decreased by 0.65 and 1.08 times whereas speed variability increased by 1.34 and 1.28 times during the TS1 and TS2, respectively, both indicating higher crash risk. However, decrease in TECG' by 64 % and 56 % during TS1 and TS2, respectively indicate compensatory measures to avoid risks due to sleep loss. A fractional regression model of crash potential revealed that low time-headway and higher speed variability and high time exposed to critical gap (TECG') significantly contribute to higher CPI values indicating higher safety risk. Other covariates such as sleep duration, professional driving experience and history of traffic violations were also associated with safety indicators and CPI, however no significant effects of age were noticed in the study. The study findings present the safety indicators sensitive to rear-end crashes specifically under PSD conditions, which can be used in designing collisions avoidance systems and strategies to improve overall traffic safety.

Assessment pedestrian crossing safety using vehicle-pedestrian interaction data through two different approaches: Fixed videography (FV) vs In-Motion Videography (IMV)

A significant portion of pedestrian accidents occurs in the outskirts areas due to the high vehicle speed and lack of safety facilities for pedestrians. Behavioral study on drivers and pedestrians is the key to better understand the causes of pedestrian accidents in order to develop safety models. Despite numerous studies on pedestrian safety based on various roads, outskirt areas have not been considered. Hence, the present study focuses on evaluating the safety of pedestrian crossing in urban and outskirt areas and to determine the differences of drivers and pedestrians' behaviors between these areas through data based on fixed videography (FV) and in-motion videography (IMV). These approaches may lead to an exact analysis of the behavioral differences of road users behaviors from the perspective of pedestrians (FV data) and drivers (IMV data) in urban and outskirts roads. Accordingly, behavioral studies were conducted at urban and outskirts sites through FV as well as IMV using the behavior of 29 participants in the same roads in Babol city, Iran. The gap acceptance model using linear regression and pedestrian crossing probability model using logistic regression for both approaches showed similarity on results in both urban and outskirts roads. Furthermore, behaviors of pedestrians crossing and drivers' yielding on urban and outskirts roads were very similar. Vehicle speed, the distance of vehicle to pedestrian at the possible collision point, size of pedestrian groups, and waiting time before crossing were the most important behavioral differences of pedestrian for choosing a gap acceptance and probability of crossing on various sites through two different approaches. The inference of the models obtained in this study will lead to a better understanding of the behavior of road users for studies on advanced driving assistance systems (ADAS).

An FPGA-Based Neuro-Fuzzy Sensor for Personalized Driving Assistance

Advanced driving-assistance systems (ADAS) are intended to automatize driver tasks, as well as improve driving and vehicle safety. This work proposes an intelligent neuro-fuzzy sensor for driving style (DS) recognition, suitable for ADAS enhancement. The development of the driving style intelligent sensor uses naturalistic driving data from the SHRP2 study, which includes data from a CAN bus, inertial measurement unit, and front radar. The system has been successfully implemented using a field-programmable gate array (FPGA) device of the Xilinx Zynq programmable system-on-chip (PSoC). It can mimic the typical timing parameters of a group of drivers as well as tune these typical parameters to model individual DSs. The neuro-fuzzy intelligent sensor provides high-speed real-time active ADAS implementation and is able to personalize its behavior into safe margins without driver intervention. In particular, the personalization procedure of the time headway (THW) parameter for an ACC in steady car following was developed, achieving a performance of 0.53 microseconds. This performance fulfilled the requirements of cutting-edge active ADAS specifications.

Practice makes better - Learning effects of driving with a multi-stage collision warning

Advanced driver assistance systems like (forward) collision warnings can increase traffic safety. As safety-critical situations (especially in urban traffic) can be diverse, integrated adaptive systems (such as multi-stage warnings) need to be developed and examined in a variety of use cases over time instead of the more common approach of testing only one-time effectiveness in the most relevant use case. Thus, this driving simulator experiment investigated a multi-stage collision warning in partially repetitive trials (T) of various safety-critical situations (scenarios confronting drivers with hazards in form of pedestrians, obstacles or preceding vehicles). Its output adapted according to the drivers' behavior in two warning stages (W1 - warning for moderate deceleration in less critical situations; W2 - urgent warning for strong, fast braking in more critical situations). To analyze how much drivers benefit from the assistance when allowed practice with it, the driving behavior and subjective ratings of 24 participants were measured over four trials. They comprised a baseline without assistance (T1) and three further trials with assistance - a learning phase repeating the scenarios from T1 twice (T2 + T3) and a concluding transfer drive with new scenarios (T4). As expected, the situation criticality in the urgent warning (W2) scenarios was rated higher than in the warning (W1) scenarios. While the brake reaction time differed more between the W1 scenarios, the applied brake force differed more between the W2 scenarios. However, the scenario factor often interacted with the trial factor. Since in later warning stages reaction time reductions become finite, the reaction strength gains importance. Overall the drivers benefited from the assistance. Both warning stages led to faster brake reactions (of similar strength) in all three assisted trials compared to the baseline, which additionally improved successively over time (T1-T3, T1 vs. T4, T2 vs. T4). Moreover, the drivers applied the gained knowledge from the learning phase to various new situations (transfer: faster brake reactions in T4 compared to T1 or T2). The well accepted and positively rated (helpful and understandable) two-stage collision warning can thus be recommended as it facilitates accident mitigation by earlier decelerations. Practice with advanced driver assistance systems (even in driving simulators) should be endorsed to maximize their benefits for traffic safety and accident prevention.

Pleasure in using adaptive cruise control: A questionnaire study in The Netherlands

OBJECTIVE

Adaptive cruise control (ACC), a technology that allows for automated car following, is becoming increasingly prevalent. Previous surveys have shown that drivers generally regard ACC as pleasant but that they have to intervene when the ACC reaches its operational limits. The former research has been mostly concerned with specific car brands and does not fully reflect the diversity of ACC types in traffic today. The objective of the present research was to establish the determinants of pleasure in using ACC.

METHODS

A 55-item online questionnaire was completed by Dutch users of diverse ACC systems.

RESULTS

Respondents (N = 182) rated their ACC highly, with a mean score of 8.0 on a scale from 1 (extraordinarily negative) to 10 (extraordinarily positive) and were most pleased with ACC on high-speed roads and in low-density traffic. Moreover, the findings point to specific operational limits such as associated with cut-in situations. Pleasure was greater for the types of ACC that are able to decelerate to a full stop, according to 48% of our sample. An analysis of the free-response items indicated that respondents who were displeased with ACC mentioned its occasional clumsiness and the dangerous situations it may evoke, whereas those who were pleased with ACC valued the complementarity of human and machine and emphasized the roles of responsibility and experience in using ACC.

CONCLUSION

Pleasure in using ACC is a function of both technological advances and human factors.

Does haptic steering guidance instigate speeding? A driving simulator study into causes and remedies

An important issue in road traffic safety is that drivers show adverse behavioral adaptation (BA) to driver assistance systems. Haptic steering guidance is an upcoming assistance system which facilitates lane-keeping performance while keeping drivers in the loop, and which may be particularly prone to BA. Thus far, experiments on haptic steering guidance have measured driver performance while the vehicle speed was kept constant. The aim of the present driving simulator study was to examine whether haptic steering guidance causes BA in the form of speeding, and to evaluate two types of haptic steering guidance designed not to suffer from BA. Twenty-four participants drove a 1.8m wide car for 13.9km on a curved road, with cones demarcating a single 2.2m narrow lane. Participants completed four conditions in a counterbalanced design: no guidance (Manual), continuous haptic guidance (Cont), continuous guidance that linearly reduced feedback gains from full guidance at 125km/h towards manual control at 130km/h and above (ContRF), and haptic guidance provided only when the predicted lateral position was outside a lateral bandwidth (Band). Participants were familiarized with each condition prior to the experimental runs and were instructed to drive as they normally would while minimizing the number of cone hits. Compared to Manual, the Cont condition yielded a significantly higher driving speed (on average by 7km/h), whereas ContRF and Band did not. All three guidance conditions yielded better lane-keeping performance than Manual, whereas Cont and ContRF yielded lower self-reported workload than Manual. In conclusion, continuous steering guidance entices drivers to increase their speed, thereby diminishing its potential safety benefits. It is possible to prevent BA while retaining safety benefits by making a design adjustment either in lateral (Band) or in longitudinal (ContRF) direction.

Use, perceptions, and benefits of automotive technologies among aging drivers

Advanced in-vehicle technologies have been proposed as a potential way to keep older adults driving for as long as they can safely do so, by taking into account the common declines in functional abilities experienced by older adults. The purpose of this report was to synthesize the knowledge about older drivers and advanced in-vehicle technologies, focusing on three areas: use (how older drivers use these technologies), perception (what they think about the technologies), and outcomes (the safety and/or comfort benefits of the technologies). Twelve technologies were selected for review and grouped into three categories: crash avoidance systems (lane departure warning, curve speed warning, forward collision warning, blind spot warning, parking assistance); in-vehicle information systems (navigation assistance, intelligent speed adaptation); and other systems (adaptive cruise control, automatic crash notification, night vision enhancement, adaptive headlight, voice activated control). A comprehensive and systematic search was conducted for each technology to collect related publications. 271 articles were included into the final review. Research findings for each of the 12 technologies are synthesized in relation to how older adults use and think about the technologies as well as potential benefits. These results are presented separately for each technology. Can advanced in-vehicle technologies help extend the period over which an older adult can drive safely? This report answers this question with an optimistic "yes." Some of the technologies reviewed in this report have been shown to help older drivers avoid crashes, improve the ease and comfort of driving, and travel to places and at times that they might normally avoid.

A quarter of a century of the DBQ: some supplementary notes on its validity with regard to accidents

This article synthesises the latest information on the relationship between the Driver Behaviour Questionnaire (DBQ) and accidents. We show by means of computer simulation that correlations with accidents are necessarily small because accidents are rare events. An updated meta-analysis on the zero-order correlations between the DBQ and self-reported accidents yielded an overall r of .13 (fixed-effect and random-effects models) for violations (57,480 participants; 67 samples) and .09 (fixed-effect and random-effects models) for errors (66,028 participants; 56 samples). An analysis of a previously published DBQ dataset (975 participants) showed that by aggregating across four measurement occasions, the correlation coefficient with self-reported accidents increased from .14 to .24 for violations and from .11 to .19 for errors. Our meta-analysis also showed that DBQ violations (r = .24; 6353 participants; 20 samples) but not DBQ errors (r = - .08; 1086 participants; 16 samples) correlated with recorded vehicle speed. Practitioner Summary: The DBQ is probably the most widely used self-report questionnaire in driver behaviour research. This study shows that DBQ violations and errors correlate moderately with self-reported traffic accidents.