Valid representation of a highly dynamic collision avoidance scenario in a driving simulator

In Germany the second-most frequent accidents in road traffic are rear-end collisions. For this reason rear-end collisions are quite important for accident research and the development of driving safety systems. To examine the functionality and to design the human–machine-interface of new driving safety systems, especially in the early development phase, subject tests are necessary. Because of the great hazard potential of such safety critical scenarios for the test persons, they are often conducted in a driving simulator (DS). Accordingly, validity is an important qualification to ensure that the findings collected in a simulated test environment can be directly transferred to the real world.This paper regards the question of driving behavior validity of DS in critical situations. There are hardly any validation studies which analyze the driving behavior in a specific collision avoidance situation.The validation study described in this paper aims to evaluate the behavioral validity of a fixed-base simulator in a collision avoidance situation. For this reason a field study from 2007 was replicated in a fixed-base simulator environment.The main questions of this validation study were if the driver can notice an active hazard braking system and if the driving behavior in a static simulator can be valid in such a critical situation.The key finding of the study states that there is no driving behavior validity in a static driving simulator for the tested dynamic scenario. The missing vestibular feedback causes a different behavior of the participants in field and simulator. The resulting absence of comparability leads to non-valid performance indicators. But these indicators are key parameters for analyzing the function and acceptance of active braking systems. So the question arises, which motion performance does a motion base have to provide in order to achieve valid acceleration simulation of such a highly dynamic collision avoidance scenario. The DS’s performance is measured in workspace, velocity and acceleration.     

Impact of time pressure on acceleration behavior and crossing decision at the onset of yellow signal

This study investigates acceleration behavior and crossing decision of the drivers under increasing time pressure driving conditions. A typical urban route was designed in a fixed-base driving simulator consisting of four signalized intersections with varying time to stop line (4 s and 6 s) and maneuver type (right-turn and go-through). 97 participants’ data were obtained under No Time Pressure (NTP), Low Time Pressure (LTP), and High Time Pressure (HTP) driving conditions. The acceleration behavior was examined at the onset of yellow signal in four ways: continuous deceleration, acceleration-deceleration, deceleration-acceleration, and continuous acceleration. A random forest model was used to build an acceleration behavior prediction model for identifying the significant explanatory variables based on variable importance ranking. Further, a Mixed Effects Multinomial Logit (MEML) model was developed using the explanatory variables obtained from a random forest model. Additionally, a generalized linear mixed model was incorporated for estimating the likelihood of crossing an intersection by considering all the explanatory variables. A MEML model result revealed that the odds of adopting acceleration-deceleration, deceleration-acceleration, and continuous acceleration instead of continuous deceleration increased by 63 %, 123 %, and 77 %, respectively under HTP driving conditions. Moreover, the likelihood of crossing a signalized intersection increased by 2.73 times and 4.26 times when the drivers were under LTP and HTP driving conditions, respectively as compared to NTP driving condition. Apart from this, time to stop line (reference: 6 s) and age showed negative association with crossing probability. Overall, the findings from this study revealed that drivers altered their acceleration behavior for executing risky driving decisions under increasing time pressure driving conditions.     

The behavioral validity of dual-task driving performance in fixed and moving base driving simulators

Next generation automotive hardware and user interfaces are increasingly pre-tested in driving simulators. What are the potential limitations of such simulations? We determined the relative and absolute validity of five different driving simulators at the Daimler AG by evaluating five functions of an in-vehicle system based on the guideline of the Alliance of Automobile Manufacturers (2006). The simulations were compared to on-road driving. We hypothesized that not only simulator characteristics, but also user characteristics, such as simulator sickness, gender, or age, influence behavioral validity. Even though relating simulator characteristics and user characteristics to driving performance across different driving simulators and driving tasks is difficult, our results are surprisingly in line with the current body of research. We demonstrate the usefulness of all simulators on a relative and partially on an absolute level with moving-base simulators being preferable to fixed-base simulators. As hypothesized, we showed that simulator sickness was significantly associated with impaired performance. In the fixed-base simulators, we found a significant interaction between age and gender, which we could not find in moving-base simulators and in the on-road study. Explanations for our findings and practical implications are discussed.     

Personality factors are associated with simulated driving outcomes across the driving lifespan

Research has shown that personality factors are related to driving safety. However, the majority of existing studies rely on self-report measures of driving behaviour and sample drivers from limited age ranges. This study sought to examine the relationship between personality and objective driving outcomes as assessed by a driving simulator in a sample of young, mid-aged, and older adults. A total of 114 active drivers completed personality questionnaires as well as a simulated driving assessment protocol. The results showed that: (1) Extraversion and neuroticism were significantly associated with driving simulator performance; (2) conscientiousness was significantly associated with driving performance among middle-aged adults; (3) sensation seeking was an important personality factor primarily for young drivers and was positively correlated with driving speed in the simulator. These results provide further support for the link between personality factors and driving performance, and suggest certain directions for future research.     

The relationship between simulator sickness and driving performance in a high-fidelity simulator

Driving simulators are highly valuable tools for various applications such as research, training, and rehabilitation. However, they are also known to cause simulator sickness, a special form of traditional motion sickness. Common side effects of simulator sickness include nausea, headache, dizziness, eye-strain, and/or disorientation, all symptoms which may negatively impact driving performance. The goal of the present study was to investigate the relationship between simulator sickness and driving performance obtained in a high-fidelity driving simulator. Twenty-one healthy participants were engaged in a simulated driving task containing rural, city, and highway sections for approx. 25 min. Participants were asked to drive naturally while obeying traffic rules and completing common driving maneuvers (including reactions to sudden events). Driving performance was evaluated based on various driving measures, such as lane positioning, speed measures, following distance, or the number of steering reversals. Simulator sickness was measured before, during, and after the simulated drive using a combination of the Simulator Sickness Questionnaire and the Fast Motion Sickness scale. Overall, correlations between the level of simulator sickness and driving performance measures were low to moderate (r’s from -0.37 to 0.40) and were not significant. Additionally, participants who reported higher levels of simulator sickness did not differ with regards to their driving performance from those who reported lower simulator sickness scores. Our results suggest that the presence of simulator sickness is not strongly related to performance in a driving simulator.     

A validation study of driving errors using a driving simulator

BackgroundDriving simulators have become an important research tool in road safety. They provide a safer environment to test driving performance and have the capacity to manipulate and control situations that are not possible on-road.AimTo validate a laboratory-based driving simulator in measuring on-road driving performance by type and mean driving errors.MethodsParticipants were instructed to drive a selected route on-road. The same route was programmed in the driving simulator using the UC/Win-road software. All participants completed a background questionnaire. On-road driving behaviours of participants and driving behaviours in the simulator were assessed by an occupational therapist and two trained researchers using an assessment form. Interclass correlations were calculated to assess the inter-rater agreement between the researchers on driving behaviours. Paired t-tests were used to assess differences in driving performance between the simulator and on-road assessments.ResultsA convenience sample of 47 drivers aged 18–69 years who held a current Western Australian class C licence (passenger vehicle) were recruited into the study. The mean age was 34.80 years (SD: 13.21) with twenty-six males (55.32%) and 21 females (44.68%) completing the study. There was no statistical difference between the on-road assessment and the driving simulator for mirror checking, left, right and forward observations, speed at intersections, maintaining speed, obeying traffic lights and stop signs.ConclusionThe preliminary results provide early support for the relative validity of the driving simulator which may be used for a variety of road safety outcomes with reduced risk of harm to participants.     

Young adult drivers' sensitivity to changes in speed and driving mode in a simple vehicle simulator

The study was done to check replication of changes in sensitivity with a simple simulator as had been obtained in an experiment using the real road situation. Another purpose was to control simulator sickness which could have confounded data from testing with a simulator or in actual driving. Sensitivity of the drivers (72 healthy young adults, M age = 24 yr., SD = 5) while performing the driving task was measured in terms of subjective ratings of simulator sickness and affect, and physiological measures (i.e., galvanic skin responses and skin temperature) at different driving speeds and in driving mode conditions, using a simple vehicle simulator. Analysis showed measures of drivers' state, including simulator sickness, physiological indices, and subjective reports, increased with driving speed (30 --> 90 -->120 km/hr.) and driving mode change from the regular speed to sudden increasing to sudden decreasing speeds. Particularly, the results suggest that the increased autonomic nervous activation induces increase of rated simulator sickness. Based upon the same tendency in change of the simulator sickness and physiological state with driving speed and driving mode conditions, it was concluded that, if the results obtained from the simulator experiment can be generalized to the real situation, the simulator sickness must be considered a confounding factor. The results also suggest that the changes in human sensitivity are dependent upon aspects related to speed of a vehicle and driving mode.     

Drivers’ driving style and their take-over-control judgment

Advanced driver assistance systems (ADASs), which help a driver drive a car safely and easily (e.g., warning alerts, steering control, and brake/acceleration pedal operation), have increased in popularity. However, such systems have not yet been perfected. Sometimes, humans must take over control from the systems; otherwise, they can cause an accident. In this study, we focused on one of the ADASs, adaptive cruise control (ACC), which automatically maintains a selected distance from the preceding car, and investigated individual differences in take-over-control judgment and related factors. The candidate factors included driver’s manual driving style, driving performance without the ACC, and the usability evaluation of ACC. Ten participants repeated the short, strictly controlled trials in a driving simulator (DS), with a varying value of only one parameter (deceleration of the preceding car) affecting the need for intervention. First, we confirmed that the participants made the judgment based on the dangerousness of the situation and that there were individual differences in the take-over-control judgments. Some participants intervened in the ACC control in less dangerous trials, whereas other participants did not, even if their own car got very close to the preceding car. We conducted a correlation analysis and confirmed the results with the estimation of the confidence interval using a bootstrap method. As a result, we found that driving style and driving performance without ACC had a stronger relationship to the number of interventions, rather than the usability evaluation. In particular, methodical drivers, who obeyed traffic rules and manners, began to intervene in less dangerous situations. The tendency to avoid utilizing brake operations was also related to take-over-control judgment. This might be because the participants intervened by pressing the brake pedal. Our study showed that drivers’ driving style could affect the usage of ACC independently from the performance of the ACC.     

Identification of driving simulator sessions of depressed drivers: A comparison between aggregated and time-series classification

Depression has been found to significantly increase the probability of risky driving and involvement in traffic collisions. The majority of studies correlating depressive symptoms with driving, pursue to predict the differences in driving behavior if the driver has already been diagnosed. Little evidence can be found, however, on how mental and psychological disorders can be identified from driving data, and usually analyses utilize simple models and aggregated data. This study aims at utilizing microscopic data from a driving simulator to detect sessions belonging to “depressed” drivers by utilizing powerful machine learning classifiers. Driving simulator sessions from 11 older drivers with symptoms of depression and 65 healthy drivers were utilized towards that aim. Random Forests, an ensemble classifier, with proven efficiency among transportation applications, are then trained on highly disaggregated data describing the mean and standard deviation of speed and lateral or longitudinal acceleration of drivers in the simulator. The kinematic data were aggregated in 30-seconds, 1-minute and 5-minute intervals, but the corresponding time-series of the measurements were also taken into account. Furthermore, classifiers were treated with imbalanced learning techniques to address the scarcity of depressed drivers among the healthy. Time-series of mean speed and the standard deviation of longitudinal acceleration even with a duration of 30-seconds have proven to be the best predictors of driving sessions belonging to depressed drivers with a very low rate of false alarms. The results outperform previous approaches, and indicate that naturalistic driving data or deep learning could prove even more efficient in detecting depression.     

Sequential dependencies in driving

The effect of recent experience on current behavior has been studied extensively in simple laboratory tasks. We explore the nature of sequential effects in the more naturalistic setting of automobile driving. Driving is a safety-critical task in which delayed response times may have severe consequences. Using a realistic driving simulator, we find significant sequential effects in pedal-press response times that depend on the history of recent stimuli and responses. Response times are slowed up to 100 ms in particular cases, a delay that has dangerous practical consequences. Further, we observe a significant number of history-related pedal misapplications, which have recently been noted as a cause for concern in the automotive safety community. By anticipating these consequences of sequential context, driver assistance systems could mitigate the effects of performance degradations and thus critically improve driver safety.     

Effects of hands-free cellular phone conversational cognitive tasks on driving stability based on driving simulation experiment

Driver distraction due to cellular phone usage is a major contributing factor to road crashes. This study compares the effects of conversational cognitive tasks using hands-free cellular phone on driving performance under three distraction conditions: (1) no distraction (no cellular conversation), (2) normal conversation (non-emotional cellular conversation), and (3) seven-level mathematical calculations. A car-following scenario was implemented using a driving simulator. Thirty young drivers with an average age of 24.1 years maintained a constant speed and distance between the subject vehicle and a leading vehicle on the driving simulator, and then respond to the leading vehicle’s emergency stop. The driving performances were assessed by collecting and statistically analyzing several variables of maneuver stability: the drivers’ brake reaction times, driving speed fluctuation, car-following distance undulation, and car-following time-headway undulation. The results revealed that normal conversation on a hands-free cellular phone impaired driving performance. The degree of impairment caused by normal calculation was equivalent to the distraction caused by Level 3 mathematical calculations according to the seven-level calculation baseline. The calculation difficulty of Level 3 is one double-digit figure plus a single-digit figure, and non-carry addition mental arithmetic is required, e.g., 44 + 4. The results indicated that an increase in the level of complexity of the calculation task was associated with an increase in brake reaction time. The seven-level calculation-task baseline could be applied to measure additional distraction effects on driving performance for further comparison.     

Effects of visual and cognitive load in real and simulated motorway driving

As part of the HASTE European Project, effects of visual and cognitive demand on driving performance and driver state were systematically investigated by means of artificial, or surrogate, In-vehicle Information Systems (S-IVIS). The present paper reports results from simulated and real motorway driving. Data were collected in a fixed base simulator, a moving base simulator and an instrumented vehicle driven in real traffic. The data collected included speed, lane keeping performance, steering wheel movements, eye movements, physiological signals and self-reported driving performance. The results show that the effects of visual and cognitive load affect driving performance in qualitatively different ways. Visual demand led to reduced speed and increased lane keeping variation. By contrast, cognitive load did not affect speed and resulted in reduced lane keeping variation. Moreover, the cognitive load resulted in increased gaze concentration towards the road centre. Both S-IVIS had an effect on physiological signals and the drivers’ assessment of their own driving performance. The study also investigated differences between the three experimental settings (static simulator, moving base simulator and field). The results are discussed with respect to the development of a generic safety test regime for In-vehicle Information Systems.     

Comparing simulator sickness in younger and older adults during simulated driving under different multisensory conditions

Driving simulators are valuable tools for traffic safety research as they allow for systematic reproductions of challenging situations that cannot be easily tested during real-world driving. Unfortunately, simulator sickness (i.e., nausea, dizziness, etc.) is common in many driving simulators and may limit their utility. The experience of simulator sickness is thought to be related to the sensory feedback provided to the user and is also thought to be greater in older compared to younger users. Therefore, the present study investigated whether adding auditory and/or motion cues to visual inputs in a driving simulator affected simulator sickness in younger and older adults. Fifty-eight healthy younger adults (age 18–39) and 63 healthy older adults (age 65+) performed a series of simulated drives under one of four sensory conditions: (1) visual cues alone, (2) combined visual + auditory cues (engine, tire, wind sounds), (3) combined visual + motion cues (via hydraulic hexapod motion platform), or (4) a combination of all three sensory cues (visual, auditory, motion). Simulator sickness was continuously recorded while driving and up to 15 min after driving session termination. Results indicated that older adults experienced more simulator sickness than younger adults overall and that females were more likely to drop out and drove for less time compared to males. No differences between sensory conditions were observed. However, older adults needed significantly longer time to fully recover from the driving session than younger adults, particularly in the visual-only condition. Participants reported that driving in the simulator was least realistic in the visual-only condition compared to the other conditions. Our results indicate that adding auditory and/or motion cues to the visual stimulus does not guarantee a reduction of simulator sickness per se, but might accelerate the recovery process, particularly in older adults.     

Compensating for failed attention while driving

While operating a motor vehicle, drivers must pay attention to other moving vehicles and the roadside environment in order to detect and process critical information related to the driving task. Using a driving simulator, this study investigated the effects of an unexpected event on driver performance in environments of more or less clutter and under situations of high attentional load. Attentional load was manipulated by varying the number of neighboring vehicles participants tracked for lane changes. After baseline-driving behavior was established, the unexpected event occurred: a pedestrian ran into the driver’s path. Tracking-accuracy, brake initiation, swerving, and verbal report of the unexpected pedestrian were used to assess driver performance. All participants verbally reported noticing the pedestrian. However, analyses of driving behavior revealed differences in the reactions to the pedestrian: drivers braked faster and had significantly less deviation in their steering heading with a lower attentional load, and participants in low clutter environments had a larger overall change in velocity. This research advances the understanding of how drivers allocate attention between various stimuli and the trade-offs between a driver’s focus on an assigned task and external objects within the roadway environment. Moreover, the results of this research lend insight into how to construct roadway environments that encourage driver attention toward the most immediate and relevant information to reduce both vehicle-to-vehicle and vehicle-to-pedestrian interactions.     

Visual load and variability of muscle activation: Effects on reactive driving of older adults

BackgroundThe functional significance of the increase in motor output variability with increased visual information processing in older adults remains unclear. Here, we test the hypothesis that increased visual information processing increases muscle activation variability in older adults and impairs their ability to react as fast and as precisely as young adults during a simulated reactive driving task.MethodsFourteen young and sixteen older adults performed a reactive driving simulation task that required responding to unexpected brake lights of the car ahead during a simple reaction time task (low visual information processing condition) and a choice reaction time task with “no go” trials condition (high visual information processing condition). We quantified the following: 1) reactive driving performance – combination of premotor response time, motor response time, and brake force error; 2) motor output variability – brake impulse variability; 3) muscle activation variability – variability in the tibialis anterior (TA) muscle activity.ResultsThe increase in information processing exacerbated the impaired reactive driving performance in older adults. The best predictor of this impairment was the increase in brake force error. The impaired reactive driving performance was related to brake impulse variability and variability in the TA activity.ConclusionsThis study provides novel evidence that increased information processing increases muscle activation variability in older adults with detrimental consequences to their ability to perform a simulated reactive driving task.     

Modeling and prediction of driver preparations for making a right turn based on vehicle velocity and traffic conditions while approaching an intersection

This paper focuses on modeling and predicting the influence of a vehicle’s velocity and the relative position between a driver’s vehicle and a vehicle to the front or rear on the onset of driver preparations for making a right turn at an intersection. Repeated experiments were carried out on a public road to measure driver preparations, including releasing the accelerator pedal, moving the right foot to cover the brake pedal, and activating the turn signal, as well as to record vehicle velocity and the relative distances to the leading and following vehicles. Structural equation modeling was applied to estimate these relationships quantitatively. Two separate latent variables accounting for the interaction between the driver’s vehicle and leading or following vehicles, and the onset point where driving behavior changes from straight mode to preparation mode before a right turn, were introduced in the specification of the structural equation model. The model estimates and testing indicate that the proposed structural equation model represents well the relationship hypothesized by observational analysis: that the vehicle velocity and the traffic conditions surrounding the driver’s vehicle strongly influence the driver’s deceleration and more weakly influenced turn signal activation. The proposed structural equation model contributes to the prediction of the onset locations of covering the brake pedal and activating the turn signal based on the vehicle velocity and the relative distances to the front and rear vehicles when the accelerator pedal is released. The prediction accuracy is high compared with predictions in which the vehicle velocity and the headway or rear distances are not taken into account or predictions using a single regression model with one independent variable, namely driving speed. Finally, a possible new addition to in-vehicle navigation systems that detects unusual driver behavior by predicting the driver’s preparatory maneuvers is discussed.     

Performance and presence with head-movement produced motion parallax in simulated driving

Driving simulator studies can reveal relevant and valid aspects of driving behavior, but underestimation of distance and speed can negatively affect the driver’s performance, such as in performance of overtaking. One possible explanation for the underestimation of distance and speed is that two-dimensional projection of the visual scene disrupts the monocular-based illusory depth because of conflicting binocular and monocular information of depth. A possible solution might involve the strengthening of the monocular information so that the binocular information becomes less potent. In the present study, we used an advanced high-fidelity driving simulator to investigate whether adding the visual depth information of motion parallax from head movement affects sense of presence, judgment of distance and speed, and performance measures coupled with overtaking. The simulations included two types of driving scenario in which one was urban and the other was rural. The main results show no effect of this head-movement produced motion parallax on sense of presence, head movement, time to collision, distance judgment, or speed judgment. However, the results show an effect on lateral positioning. When initiating the overtaking maneuver there is a lateral positioning farther away from the road center as effect of the motion parallax in both types of scenario, which can be interpreted as indicating use of naturally occurring information that change behavior at overtaking. Nevertheless, only showing tendencies of effects, absent is any clear additional impact of this motion parallax in the simulated driving.     

Investigating the influence of time pressure on overtaking maneuvers and crash risk

The research conducted on overtaking maneuver for evaluating drivers’ safety showed adverse effects of urgency on driving performance and decision making. Therefore, a driving simulator study was designed to examine driving performance of the drivers and its implication on overtaking and crash probabilities under increasing time pressure conditions. Eighty-eight participants data were analyzed in the current study. Three different time pressure conditions: No Time Pressure (NTP), Low Time Pressure (LTP), and High Time Pressure (HTP) were considered for analyzing driving performance of the drivers while executing overtaking maneuvers. The driving performance was assessed using minimum time-to-line crossing and coefficient of variation in speed to dissect the safety margin adopted by the drivers while overtaking the lead vehicle. Further, minimum time-to-line crossing and coefficient of variation in speed were considered as explanatory variables to investigate their influence on overtaking and crash probabilities. Parametric survival analysis and Generalized Linear Mixed Models (GLMM) were used to assess the driving performance, overtaking and crash probabilities. The parametric survival analysis showed that minimum time-to-line crossing reduced by 36.7% and 63.8% in LTP and HTP driving conditions, respectively. The GLMM results revealed that coefficient of variation in speed increased by 3.437% in HTP (no significant effect in LTP) as compared to NTP driving conditions. Further, the GLMM results showed that overtaking and crash probabilities decreased with increment in minimum time-to-line crossing and coefficient of variation in speed values. Additionally, it was observed that male drivers took risky decisions than female drivers. Nevertheless, the comparative analysis revealed that male drivers were less prone to crashes than female drivers. Overall, it can be inferred that the drivers take risky decisions with increment in time pressure to complete the driving task, even at the expense of their own safety which exposed them to high likelihood of crashes.     

How drivers adapt to drive in driving simulator,and what is the impact of practice scenario on the research?

Adaptation is an important precondition for validity of experiments carried out using a driving simulator. Learning how to control a simulated vehicle imposes mental load on participants, which can potentially distract them from performing the main task and bias the results of experiments. Most researchers have a practice session before the main scenario to ensure participants have adapted. However, the practice scenarios vary greatly both in duration and form. Moreover, in almost all cases, it is difficult to verify that a participant is adapted at the end of the practice session. It was previously shown that using a power curve can mathematically model the learning pattern of subjects to steering and pedal control and can also help identify adapted and non-adapted subjects at the end of practice scenarios. The same concept is implemented in the current study to evaluate the appropriateness of such a methodology for a different practice scenario. The results showed that adaptation time and learning rate did not differ significantly between male and female participants. More importantly, statistical evidence verified that adaptation to a driving simulator is task-independent. The impact of the practice scenario on the main task was also analyzed, leading to the observation that during the experiment scenario participants tend to continue to improve the sub-skills they focused on during the practice scenario. Based on the results of these adaptation sessions, recommendations are provided to improve the quality of design for the practice scenario and to minimize its impact on the experiment scenario.