Effect of multimodal takeover request issued through A-pillar LED light,earcon, speech message,and haptic seat in conditionally automated driving

The driver of a conditionally automated vehicle equivalent to level 3 of the SAE is obligated to accept a takeover request (TOR) issued by the vehicle. Considerable research has been conducted on the TOR, especially in terms of the effectiveness of multimodal methods. Therefore, in this study, the effectiveness of various multimodalities was compared and analyzed. Thirty-six volunteers were recruited to compare the effects of the multimodalities, and vehicle and physiological data were obtained using a driving simulator. Eight combinations of TOR warnings, including those implemented through LED lights on the A-pillar, earcon, speech message, or vibrations in the back support and seat pan, were analyzed to clarify the corresponding effects. When the LED lights were implemented on the A-pillar, the driver reaction was faster (p = 0.022) and steering deviation was larger (p = 0.024) than those in the case in which no LED lights were implemented. The speech message resulted in a larger steering deviation than that in the case of the earcon (p = 0.044). When vibrations were provided through the haptic seat, the reaction time (p < 0.001) was faster, and the steering deviation (p = 0.001) was larger in the presence of vibrations in the haptic seat than no vibration. An interaction effect was noted between the visual and auditory modalities; notably, the earcon resulted in a small steering deviation and skin conductance response amplitude (SCR amplitude) when implemented with LED lights on the A-pillar, whereas the speech message led to a small steering deviation and SCR amplitude without the LED lights. In the design of a multimodal warning to be used to issue a TOR, the effects of each individual modality and corresponding interaction effects must be considered. These effects must be evaluated through application to various takeover situations.     

Evidence that implementation intentions reduce drivers’ use of mobile phones while driving

Implementation intentions are IF-THEN plans that have the potential to reduce mobile phone use while driving and thus contribute towards the prevention of road traffic crashes. We tested whether an intervention, designed to promote the formation of implementation intentions, could reduce drivers’ use of mobile phones. A randomized controlled design was used. The participants (N = 136) were randomised to an implementation intention or a control condition. Self-report questionnaires were administered to all participants at both pre- and one-month post-intervention to measure the use of mobile phones while driving, goal intentions and the theoretically derived motivational pre-cursors of goal intentions (attitudes, subjective norm and perceived behavioural control). Immediately following the pre-intervention questionnaire, the participants in the implementation intention condition (n = 67) were given a volitional help sheet, which asked them to form implementation intentions by specifying target driving situations that tempted them the most to use a mobile phone and linking them with goal-directed responses that could be used to resist the temptation. The participants in the control condition (n = 69) were asked to specify target situations that tempted them the most to use a mobile phone while driving and to generally try to avoid using a mobile phone in those situations. One-month post-intervention, the participants in the implementation intention condition reported using a mobile phone less often while driving in their specified target driving situations than did the participants in the control condition. As expected, no differences were found between the conditions in the reported frequency of mobile phone use in unspecified driving situations, goal intentions or any motivational pre-cursor of goal intentions. The implementation intention intervention that was tested in this study is a potentially effective tool for reducing mobile phone use while driving in target driving situations, where behaviour-change is most needed.     

Semi-automated versus highly automated driving in critical situations caused by automation failures

The purpose of this study was to examine the effects of vehicle automation and automation failures on driving performance. Previous studies have revealed problems with driving performance in situations with automation failures and attributed this to drivers being out-of-the-loop. It was therefore hypothesized that driving performance is safer with lower than with higher levels of automation. Furthermore, it was hypothesized that driving performance would be affected by the extent of the automation failure. A moving base driving simulator was used. The design contained semi-automated and highly automated driving combined with complete, severe, and moderate deceleration failures. In total the study involved 36 participants. The results indicate that driving performance degrades when the level of automation increases. Furthermore, it is indicated that car drivers are worse at handling complete than partial deceleration failures.     

Behavioural adaptation and effectiveness of a Forward Collision Warning System depending on a secondary cognitive task

Forward Collision Warning Systems (FCWS) have been designed to enhance road safety by reducing the number of rear-end collisions. Nevertheless, little is known about how drivers adapt their behaviour over time when using this kind of system. In addition, these systems are expected to aid particularly distracted drivers. However, previous research has suggested that the effectiveness of the system could depend on the difficulty level of the secondary task. The objective of this study on driving simulator was twofold. Firstly, it consisted in evaluating the behavioural adaptation to an FCWS as well as analysing the possible consequences of driving without the system after a short period of adaptation. Secondly, it was to evaluate the effectiveness of the system according to two different difficulty levels of a cognitive secondary task. The results showed that drivers adapted their behaviour positively when the system was introduced. Nevertheless, both the effectiveness and the behavioural adaptation in the short term were dependent on the cognitive load induced by the secondary task. These findings suggest that the warning needs some attentional resources to be processed. Finally, no negative or transfer effect was observed following the removal of the system after a short period of adaptation.     

Towards a new approach to detect sleepiness: Validation of the objective sleepiness scale under simulated driving conditions

The Objective Sleepiness Scale (OSS) was developed to detect and quantify sleepiness on the basis of two direct and reliable sleepiness indicators: EEG and EOG. The present study aims to test whether the OSS can be used to detect sleepiness episodes that impair performance on driving and vigilance tasks accurately and with a good time synchronization. Forty-three healthy volunteers performed monotonous driving sessions on a simulator and the psychomotor vigilance task (PVT) in a normal sleep condition and after partial sleep deprivation. OSS reliability and time synchronization for sleepiness detection were tested on driving (standard deviation of vehicle lateral position and off-road duration) and PVT (reaction time and lapses). Inter-rater reliability of the scale was evaluated by two blinded scorers. Results show that the OSS score indicates higher sleepiness in sleep deprivation conditions (p < 0.001) and with time-on-task. Differences of performance between OSS score calculated with multiple pairwise comparisons, indicate OSS score increase when driving performance (SDLP and off-road duration) decreases (p < 0.001 for comparisons between OSS stages 0 vs 2, 0 vs 3, 1 vs 2 and p < 0.05 for 1 vs 3). Reaction time during PVT is also related to the OSS score (p < 0.05 for OSS values from 0 to 2, 0 to 3, 1 to 2 and 1 to 3). There is no proportional relation between OSS score and performance impairment, but a threshold effect between levels 1 and 2 of the scale is observed. Positive outcomes are also obtained for time synchronization of the OSS assessed on driving performance (p < 0.001 for both SDLP and off-road duration). Finally, inter-rater agreement is found to be considerable. The results allow us to consider using the Objective Sleepiness Scale as a tool for research on sleepiness.     

Transition patterns of driving style from a traditional driving environment to a connected vehicle environment: A case of an extra-long tunnel road

To provide a better understanding of individual driver’s driving style classification in a traditional and a CV environment, spatiotemporal characteristics of vehicle trajectories on a road tunnel were extracted through a driving simulator-based experiment. Speed, acceleration, and rate of acceleration changes are selected as clustering indexes. The dynamic time warping and k-means clustering were adopted to classify participants into different risk level groups. To assess the driver behavior benefits in a CV environment, an indicator BI (behavior indicator, BI) was defined based on the standard deviation of speed, the standard deviation of acceleration, and the standard deviation of the rate of acceleration change. Then, the index BI of each driver was calculated. Furthermore, this paper explored driving style classification, not in terms of traditional driving environment, but rather the transition patterns from a traditional driving environment to a CV environment. The results revealed that inside a long tunnel, 80 % of drivers benefited from a CV environment. Moreover, drivers might need training before using a CV system, especially female drivers who have low driving mileage. In addition, the results showed that the driving style of 69 % of the drivers’ transferred from a high risk-level to a low risk-level when driving in a CV environment. The study results can be expected to improve driving training education programs and also to provide a valuable reference for developing individual in-vehicle human-machine interface projects and other proactive safety countermeasures.     

Effects of environmental,vehicle and human factors on comfort in partially automated driving: A scenario-based study

Although it is key to improving acceptability, there is sparse scientific literature on the experience of humans as passengers in partially automated cars. The present study therefore investigated the influence of road type, weather conditions, traffic congestion level, vehicle speed, and human factors (e.g., trust in automated cars) on passenger comfort in an automated car classified as Level 3 according to the Society of Automotive Engineers (SAE). Participants were exposed to scenarios in which a character is driven by an SAE Level 3 automated car in different combinations of conditions (e.g., highway × heavy rain × very congested traffic × vehicle following prescribed speed). They were asked to rate their perceived comfort as if they were the protagonist. Results showed that comfort was negatively affected by driving in downtown (vs. highway), heavy rain, and congested traffic. Interaction analyses showed that reducing the speed of the vehicle improved comfort in these two last conditions, considered either individually or in combination. Cluster analysis revealed four profiles: trusting in automation, averse to speed reduction, risk averse, and mistrusting automation. These profiles were all influenced differently by the driving conditions, and corresponded to varying levels of trust in automated cars. This study suggests that optimizing comfort in automated cars should take account of both driving conditions and human profiles.     

Other road users’ adaptations to increase safety in response to older drivers’ behaviour

Changes in physical and cognitive abilities not only challenge the driving ability of older adults, in some situations age-related changes in driving behaviour require other road users to adapt their behaviour to maintain a safe traffic situation. In this study, we aimed to map age-related differences in driving behaviour and assess the impact on other road users. A group younger and a group older adults drove four different routes containing challenging situations (e.g., merging into motorway traffic) in a driving simulator while measures of driving behaviour were collected. Other road users’ deceleration responses to the driver’s behaviour were also collected as a measure of behavioural adaptation. Our results showed similar driving performance between young and older drivers when task complexity was low, but reduced performance in older drivers when tasks requirements increased. Lower driving speed and longer waiting times that were observed in older drivers can be interpreted as compensatory behaviour aimed at creating more time to lower task requirements. Crucially, in a non-time critical situation this compensatory behaviour was found to be successful, however in a time-critical situation (merging onto a motorway) this strategy had negative side effects because other road users had to decelerate in order to keep a safe distance. Our results show the importance of anticipation and adaptation by other road users for the success of older driver’s strategies and traffic safety.     

Behavior model and guidance strategies of the crossing behavior at unsignalized intersections in the connected vehicle environment

The connected vehicle environment is considered to be a disruptive technology that reconstructs the way people travel and road transport, and more importantly, ensures traffic safety. This study aims on investigating drivers’ interactive behavior at an unsignalized intersection in the connected vehicle environment. Specifically, a simplified iterative behavior model was established to predict the potential conflicting vehicles’ behavior. Furthermore, based on principles of safety, efficiency, and comfort, the guidance strategies were proposed to help the subject vehicles cross intersections. A multi-user driving simulator experiment was carried out and 48 participants were divided into 24 pairs to complete the test. The simplified iterative behavior model was constructed based on the dynamic interaction of each participant pair as they approached the intersection from straight-crossing directions. The comparison results showed the behavior model was an effective microscopic simulation tool for vehicles at unsignalized intersections with high accuracy and good applicability. Then, the guidance strategies under different compliance rates (baseline, 50 % compliance, 100% compliance) were evaluated based on three indexes i.e. standard deviation of speed (SDS), duration of crossing the intersection (DCI), and time exposed post-encroachment-time (TEP). The numerical simulation results showed that the guidance strategies could effectively improve the safety and efficiency of drivers crossing the intersection under both 50% and 100% compliance rates. Besides, with the increase of compliance rate, the comfort level also increased evidently. This study can provide theoretical and algorithmic references for microscopic simulation and guidance strategy at unsignalized intersections in the connected vehicle environment.     

Driving self-regulation and risky driving outcomes

Self-regulation has been associated with risky driving outcomes in the past but there are no available measures to assess driving-specific self-regulatory capacity. The present study assessed the association of a newly developed driving self-regulation measure with driving violations, errors, and lapses. Overall, 330 UK drivers completed measures of risky driving outcomes, driving anger, trait impulsivity, sensation seeking, normlessness, domain-general trait self-regulation plus a new unidimensional measure of Driving Self-Regulation Questionnaire (DSRQ-16). Bivariate correlation analysis indicated that the DSRQ-16 showed expected associations with both driving-related outcomes and factors, as well as with impulsivity traits and general self-regulation. Bootstrapped hierarchical linear regression models showed that the DSRQ-16 was significantly associated with driving violations, errors, and lapses after controlling for the effects of other relevant predictors. This is the first study to demonstrate the association of driving-specific self-regulation with risky driving behaviour, driving anger, impulsivity and related personality traits. Driving-specific self-regulation may present a novel target for road safety interventions, as well as a theoretically relevant component of models of risky driving behaviour.