Takeover Time Research Articles

Identifying interaction types and functionality for automated vehicle virtual assistants: An exploratory study using speech acts cluster analysis

Onboard virtual assistants with the ability to converse with users are gaining favour in supporting effective human-machine interaction to meet safe standards of operation in automated vehicles (AVs). Previous studies have highlighted the need to communicate situation information to effectively support the transfer of control and responsibility of the driving task. This study explores ‘interaction types’ used for this complex human-machine transaction, by analysing how situation information is conveyed and reciprocated during a transfer of control scenario. Two human drivers alternated control in a bespoke, dual controlled driving simulator with the transfer of control being entirely reliant on verbal communication. Handover dialogues were coded based on speech-act classifications, and a cluster analysis was conducted. Four interaction types were identified for both virtual assistants (i.e., agent handing over control) - Supervisor, Information Desk, Interrogator and Converser, and drivers (i.e., agent taking control) - Coordinator, Perceiver, Inquirer and Silent Receiver. Each interaction type provides a framework of characteristics that can be used to define driver requirements and implemented in the design of future virtual assistants to support the driver in maintaining and rebuilding timely situation awareness, whilst ensuring a positive user experience. This study also provides additional insight into the role of dialogue turns and takeover time and provides recommendations for future virtual assistant designs in AVs.

Tactile Displays: The Effects of Location and Intensity on Automated Vehicle Takeover Performance

This study investigated the effects of tactile display location and intensity on automated vehicle takeover performance. Twenty participants completed simulated driving tasks in a Level 3 automated vehicle and responded to takeover requests presented at different locations (seat back, seat pan, seat belt, and wrist) with different intensity levels. Results showed that higher intensity tactile displays were associated with faster takeover times, and placing the display on the wrist was correlated with shorter takeover times compared to the seat pan in lower intensity. However, there were no significant effects on maximum resulting jerk. These findings suggest that incorporating tactile displays in automated vehicles as takeover requests is promising, particularly when placed on the wrist and presented at higher intensities. This research contributes to the development of human-machine interfaces that can better assist drivers during takeovers and improve the safety of automated vehicles.

Drivers’ trust & reliance on automated braking at intersections: Effects of automation braking profile & in-vehicle displays

An online, video-based, within-subjects experiment was conducted to investigate the effects of three different automation braking profiles and three in-vehicle display designs on drivers’ trust and reliance on automated braking systems. 36 participants watched videos of automated braking at intersections (with different braking times and displays) and rated their trust in the automation and comfort with the automation’s braking time. They then watched the same videos and were asked to press the space bar and pause the video if they believed they should take over from the automation and brake. The braking profile had a significant effect on the trust and comfort ratings, while both the braking profile and display had a significant effect on the takeover time. The findings provide insights on drivers’ trust and reliance on automation that is designed to avoid lateral collision hazards at intersections.

Predicting Driver Takeover Decisions in Conditionally Automated Vehicles with a Gaze-based Deep Learning Model

This study proposes a novel methodology for modeling driver takeover behavior in conditionally automated vehicles (AVs) when exiting a freeway using deep learning (DL) network architectures. While previous research has focused on modeling takeover time in emergency scenarios, which require quick responses, these models may not be applicable to scheduled, non-time-critical takeovers. In such situations, drivers may employ varying strategies and take longer to resume control of the vehicle when there is no time pressure. To address this problem, a deep learning architecture based on a convolutional neural network (CNN) was implemented to predict drivers’ takeover behaviors in scheduled takeovers. The model was trained on drivers’ driving data and eye gaze with varying time windows, facilitating an analysis of drivers’ takeover decisions to various takeover request designs. The model achieved good performance metrics, with an F1 Score of 0.993, a recall of 0.996, and a precision of 0.991. The application of these models holds substantial potential for refining the design of the human-machine interface, specifically in calibrating the takeover request (ToR) lead time, thereby promoting safe freeway exiting takeovers in conditionally AVs.

Measuring takeover performance in different driving scenarios: Considering the influence of non-driving-related tasks and takeover request lead time

Conditional automation systems (SAE Level 3) allow the driver to perform a secondary task, but the driver must regain control of the vehicle after the takeover request (ToR). Therefore, it is important to understand the factors that influence the driver’s takeover performance. Previous studies have analyzed the variables including ToR modalities, ToR lead time, non-driving-related (NDR) task, driving experience and age. The influence of driving scenario on driver takeover performance and whether this influence varies the effect of the ToR time and NDR task remained unanswered. This study designed a 2(ToR time) × 2(NDR task) × 4 (driving scenario) factorial measure to investigate the effect of these factors on automated vehicle takeover performance. Forty-two participants drove in a simulated SAE Level 3 vehicle and performed a series of takeover tasks. The results showed that the ToR time, NDR task, driving scenario and ToR time × driving scenario had a significant effect on takeover behavior. In particular, the secondary work state task was correlated with the increased takeover and control time. The complexity of the driving scenario indicated a marginally significant influence on the quality of the takeover. A shorter takeover time and a smaller lateral offset were associated with a lower urgency of the driving scenario. The ToR time for the fog-cluster scenario showed an opposite effect to the takeover performance compared to the other three scenarios. Findings from this study can provide implications on the effect of ToR time desired and enabled by different scenarios.

자율주행 환경에서 비운전과제 유형과 도로 혼잡도에 따른 청년 및 고령운전자의 제어권 인수와 운전수행 차이

The take-over time (TOT) and driving performances of young and old drivers according to the types of non-driving related tasks (NDRTs) and road complexity were compared by using 3-level automated driving simulation. NDRTs were divided into four types according to the stimulus modality and response modality: visual stimulus-verbal response(V-V), visual stimulus-manual response(V-M), visual/auditory stimulus-manual response(AV-M), and auditory stimulus-verbal response(A-V). Control condition where NDRT was not performed was also included. The participant group was divided into a young driver group and an elderly driver group, and the road condition was divided into a no-traffic condition and a heavy traffic condition based on the surrounding traffic volume. The main results of this study are as follows. First, TOT was the slowest in the visual stimulus-manual response condition, but the fastest in the auditory stimulus-verbal response condition. In particular, this tendency was more prominent in the elderly driver group than in the younger driver group and in the heavy traffic condition than in the no-traffic condition. Second, the main effect of the NDRT type was not significant in the driving performance measurements including steering wheel manipualtion variability and brake input after taking over control, but greater steering wheel manipulation variability and brake input were observed in the elderly driver group compared to the young driver group and in the heavy traffic condition compared to the no-traffic condition. These results suggest that the driver's age, NDRT type, and road conditions can have an integrated effect on TOT and driving performance after taking over control in autonomous driving situations.

Playing Games Guiding Attention Improves Situation Awareness and Takeover Quality during Automated Driving

During SAE L3 automated driving, drivers engage in non-driving related tasks (NTRDs), but NTRDs impair situation awareness (SA) and takeover quality. This study aimed to find an NTRD that is interesting but has a minimal negative impact on SA. Given the strong impact of games on cognition, we studied how games affected SA and takeover performance. In the game survey, we investigated the impact of four type of games on driving SA. The results showed that there were some games positively affected SA (situational game). In the formal experiment, we used a realistic simulated driving environment to investigate the effect of games on SA and takeover performance. A 3 (task type: situational game/non-situational game/monitoring) × 4 (attention duration, time between monitoring requests and takeover requests: 4 s/6 s/8 s/10 s) mixed design was used. Results showed that drivers who engaged in the situational game had higher SA, better takeover performance, and required less attention duration than drivers who engaged in the non-situational game. Both groups showed faster takeover time than the monitoring group, indicating higher arousal. Our research results indicate that playing situational games can improve SA and takeover quality in automated driving.

Transitioning to multi-dimensional estimation of visual distraction and its safety effects under automated driving: A spatiotemporal and directional estimation approach

Traditional methodologies for measuring visual distraction have been limiting in their approaches, treating distraction as a one-dimensional variable. This has been accomplished either by categorizing distraction as a binary variable or using surrogate measurements, such as reaction time in response to non-driving related tasks, which follow the procedures of measuring distraction from psychology. Furthermore, as human-vehicle interaction (HVI) under automated driving has the potential to provide safety information and bring visual distraction simultaneously, there lacks an investigation on the quantitative relationship between distraction and safety due to the restrained methodologies in measuring visual distraction. As HVI-induced driver distraction plays a critical role in determining safe driving under automated driving, a methodology is highly needed to comprehensively measure the visual distraction under automated driving so that its impact on safety can be further investigated. Therefore, sticking to the definition of visual distraction, this research aims to (1) improve the existing methodology in measuring HVI-induced driver distraction under automated driving by focusing on visual distraction and mathematically describing it from spatiotemporal and directional dimensions; (2) theoretically investigate how the quantified visual distraction influences driving safety and (3) quantify the relationship between distraction and safety with introducing the “distraction-safety” ratio. Drivers’ fixation behaviors are used in quantifying visual distraction, which is measured by spatiotemporal and directional relationships between drivers’ visual attention and the attention that indicates “zero distraction”. Three newly added performance measures in quantifying HVI-induced visual distraction are real-time magnitudes, real-time directions, and intensities (cumulation of magnitudes over time). To validate the proposed methods, a verification study was conducted by recruiting drivers to test automated driving under Level 3 automation. Drivers are required to wear an eye-tracker and go through two scenarios interacting with jaywalkers where takeover actions are needed with two takeover warnings (“visual-only” and “visual & audible”). Per past studies, takeover time was measured in representing distraction level. As a result, this study confirms the validity of the proposed methods by revealing the significant and positive correlations between the measured distraction intensity and the takeover time. Discussions of the quantified visual distraction from the spatiotemporal and directional perspective further enhance the understanding of HVI-induced driver distraction under automated driving through multiple dimensions. Furthermore, this research reveals how distracted driving affects safety under automated driving through the takeover performance. Thresholds under visual distraction magnitudes and degrees that lead to traffic conflicts were identified. More importantly, a “distraction-safety” ratio that quantifies the relationship between visual distraction and safety benefits is proposed. The results suggest that the “visual & audible” is more effective in significantly enhancing safety while aggregating a significantly smaller amount of visual distraction. The contribution of this research is to re-define the methodological approach of measuring visual distraction by (1) measuring distraction multi-dimensionally and (2) establishing a “distraction-safety” system in quantitatively assessing the balance of safety benefits and the HVI-induced visual distraction magnitude under automated driving environment.

Are Drivers Allowed to Sleep? Sleep Inertia Effects Drivers’ Performance after Different Sleep Durations in Automated Driving

Higher levels of automated driving may offer the possibility to sleep in the driver’s seat in the car, and it is foreseeable that drivers will voluntarily or involuntarily fall asleep when they do not need to drive. Post-sleep performance impairments due to sleep inertia, a brief period of impaired cognitive performance after waking up, is a potential safety issue when drivers need to take over and drive manually. The present study assessed whether sleep inertia has an effect on driving and cognitive performance after different sleep durations. A driving simulator study with n = 13 participants was conducted. Driving and cognitive performance were analyzed after waking up from a 10–20 min sleep, a 30–60 min sleep, and after resting without sleep. The study’s results indicate that a short sleep duration does not reliably prevent sleep inertia. After the 10–20 min sleep, cognitive performance upon waking up was decreased, but the sleep inertia impairment faded within 15 min. Although the driving parameters showed no significant difference between the conditions, participants subjectively felt more tired after both sleep durations compared to resting. The small sample size of 13 participants, tested in a within-design, may have prevented medium and small effects from becoming significant. In our study, take-over was offered without time pressure, and take-over times ranged from 3.15 min to 4.09 min after the alarm bell, with a mean value of 3.56 min in both sleeping conditions. The results suggest that daytime naps without previous sleep deprivation result in mild and short-term impairments. Further research is recommended to understand the severity of impairments caused by different intensities of sleep inertia.

Investigating effects of multimodal explanations using multiple In-vehicle displays for takeover request in conditionally automated driving

In conditionally automated driving, a human takeover transition is necessary in response to vehicle malfunction or other limitation. To mitigate the negative experience of this technological limitation, researchers have suggested providing an explanation along with the takeover request to increase drivers’ understanding and trust in an automated driving system (ADS). However, previous studies on the explanation have been limited to investigations of unimodal explanations that use either text or speech, while neglecting other various in-vehicle display options. The present study aims to investigate the effects of multimodal explanation using three in-vehicle displays—ambient light, cluster, and sound. The explanation given by each in-vehicle display was systemically manipulated into two levels (ambient light: constant color vs. variable color, cluster: no-text and no-background vs. text with colored background, sound: alert-only sound vs. speech with alert sound). To examine the effects of the explanation using these displays, we conducted mixed method research including a scenario-based experiment, questionnaires, and a follow-up interview for thematic analysis. In the experiment, eighty-two participants drove in a driving simulator and performed takeover tasks in several types of urgent situations for measuring their takeover time and error. The participants then responded to questionnaires on takeover workload and overall trust in automated vehicle. Next, fifty-six of them participated in the follow-up interview for discovering significant themes related to their experiences with the multimodal explanation. As results, we found that multimodal explanations that use two types of displays were more effective than unimodal explanations or multimodal explanations using three types of displays. Regarding the effects of each type of display, the speech-based explanation showed the best takeover performance in general. The explanation using the cluster was next most useful in performance. The ambient light explanation showed less significant to no effect in the takeover situation. We expect that these findings can contribute to theoretical understanding of driver response to multimodal explanations and to practical design of trustworthy explanations in various takeover situations.

Reliable and transparent in-vehicle agents lead to higher behavioral trust in conditionally automated driving systems.

Trust is critical for human-automation collaboration, especially under safety-critical tasks such as driving. Providing explainable information on how the automation system reaches decisions and predictions can improve system transparency, which is believed to further facilitate driver trust and user evaluation of the automated vehicles. However, what the optimal level of transparency is and how the system communicates it to calibrate drivers' trust and improve their driving performance remain uncertain. Such uncertainty becomes even more unpredictable given that the system reliability remains dynamic due to current technological limitations. To address this issue in conditionally automated vehicles, a total of 30 participants were recruited in a driving simulator study and assigned to either a low or a high system reliability condition. They experienced two driving scenarios accompanied by two types of in-vehicle agents delivering information with different transparency types: "what"-then-wait (on-demand) and "what + why" (proactive). The on-demand agent provided some information about the upcoming event and delivered more information if prompted by the driver, whereas the proactive agent provided all information at once. Results indicated that the on-demand agent was more habitable, or naturalistic, to drivers and was perceived with faster system response speed compared to the proactive agent. Drivers under the high-reliability condition complied with the takeover request (TOR) more (if the agent was on-demand) and had shorter takeover times (in both agent conditions) compared to those under the low-reliability condition. These findings inspire how the automation system can deliver information to improve system transparency while adapting to system reliability and user evaluation, which further contributes to driver trust calibration and performance correction in future automated vehicles.

Impact of non-driving related tasks while operating automated driving systems (ADS): A systematic review

Automated Driving Systems (ADS) (SAE, 2021), promise improved safety and comfort for drivers. Current technological advances have resulted in increased automation capabilities. However, with the increase in automation capabilities, there is a shift in how drivers interact with their vehicles. Drivers can now temporarily hand over the control of the driving task to ADS under certain conditions. However, with ADS in temporary control of the vehicle, drivers may choose to engage in non-driving related tasks (NDRT). The current capabilities of ADS do not allow drivers to hand over control of the driving task indefinitely. Drivers must remain aware and be ready to take back control if necessary. There is a need to better understand drivers’ performance and behaviors when driving with ADS, especially when engaged in NDRTs. This literature review, therefore, aims to understand the state of knowledge on automated vehicle systems and driver distraction. This review was conducted as per PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies found a significant increase in takeover times while engaging in NDRTs and driving with automation active. Studies also discuss a change in driver’s visual attention, with more focus given to NDRTs as compared to the front roadway. The concerning effects of increasing reaction times and decreases in visual attention can be mitigated by using interventions and studies have had success in redirecting drivers attention and reorient them to the task of driving. The review, therefore, includes a discussion of ADS and NDRT engagement and its impact on driving behaviors such as take-over times, visual attention, trust, and workload. Implications on driver safety and performance are discussed in light of this synthesis.

Predicting Driver Errors during Automated Vehicle Takeovers

The transition of control between partially automated vehicles and drivers is an important part of the operational design domain and poses unique and important design issues. One approach for enhancing the design of the transition of control mechanisms is to predict driver behavior during a takeover by analyzing his/her state before a takeover. Although there is a wealth of existing literature on modeling the prediction of driver behavior, little is known about the prediction of takeover performance (e.g., driver error) and its underlying data structure (e.g., window sizes or the inclusion of certain features). Thus, the goal of this study is to predict driver error during a takeover event using supervised machine learning algorithms for various window sizes. Three machine learning algorithms (i.e., decision tree, random forest, and support vector machine with a radial basis kernel) were applied to granular driving performance, physiological, and glance data from a driving simulator experiment examining automated vehicle driving. The results showed that a random forest algorithm with an area under the receiver operating characteristic curve of 0.72, trained on a 3 s window before the takeover time, had the best performance with regard to classifying driver error accurately. In addition, we identified the 10 most important predictors that resulted in the best error prediction performance. The results of this study could be useful in developing algorithms for driver state that could be integrated into highly automated systems and, potentially, improve the takeover process.

Impact of duration of monitoring before takeover request on takeover time with insights into eye tracking data

Safety has become the primary concern of automated driving system (ADS) in recent years. Compared with highly automated driving (L4 and above), conditionally automated driving (L3/L3+ ADS) seems to be a moderate choice, where drivers are required to respond to the takeover request (TOR) whenever necessary. It is the system’s responsibility to make sure that the takeovers would be safe at the time of issuing the TOR. To realize that, a lot of factors need to be taken into consideration. As it has been found that drivers’ eyes-on-road gazes increase slowly in the first few seconds while transferring to manual driving from automated driving and drivers’ gaze behaviors are related with situation awareness, the main aim of this study is to investigate the impact of duration of monitoring before the TOR on takeover time and whether there is a positive or negative relationship between the two. To verify these, we designed a driving simulator study where the TOR was issued 0 s, 5 s, 10 s and ≥ 15 s after the non-driving-related task has ended. Twelve scenarios were designed, and the results from 36 participants showed that there was indeed a statistically significant difference, however, the relationship was neither positive nor negative, which was close to a parabola. Analyzing results of eye movements and gaze behavior further supported this conclusion. It is therefore concluded the duration of monitoring before the TOR should neither be too short nor too long, and 5–7 s would be appropriate choices. This is desirable not only for improving takeover performance of drivers but also for improving the prediction model for predicting takeover performance of drivers that has yet to be studied, so as to improve safety, reliability and acceptance of the ADS.

Keeping in the lane! Investigating drivers’ performance handling silent vs. alerted lateral control failures in monotonous partially automated driving

This study investigated how drivers can manage the take-over when silent or alerted failure of automated lateral control occurs after monotonous hands-off driving with partial automation. Twenty-two drivers with varying levels of prior ADAS experience participated in the driving simulator experiment. The failures were injected into the driving scenario on curved road segments, accompanied by either a visual-auditory alert or no change in HMI. Results indicated that drivers could rarely maintain lane-keeping when automated steering was disabled silently, but most drivers safely managed the alerted failure situation within the ego-lane. The silent failure yielded significantly longer take-over time and generally worse lateral control quality. In contrast, poor longitudinal control performance was observed in alerted conditions due to more brake usage. An expert-based controllability assessment method was introduced to this study. The silent lateral failure situation during monotonous hands-off driving was rated as uncontrollable, while the alerted situation was basically controllable. Participants showed their preferences for the TORs, and the importance of conveying TOR reasons was also demonstrated. Relevance to industryThe results and implications of this study provided insights into the design and development of automated driving systems to prevent critical consequences. The comprehensive method of controllability assessment can benefit the automated driving system evaluation.

A Study on the Entire Take-Over Process-Based Emergency Obstacle Avoidance Behavior.

Nowadays, conditional automated driving vehicles still need drivers to take-over in the scenarios such as emergency hazard events or driving environments beyond the system's control. This study aimed to explore the changing trend of the drivers' takeover behavior under the influence of traffic density and take-over budget time for the entire take-over process in emergency obstacle avoidance scenarios. In the driving simulator, a 2 × 2 factorial design was adopted, including two traffic densities (high density and low density) and two kinds of take-over budget time (3 s and 5 s). A total of 40 drivers were recruited, and each driver was required to complete four simulation experiments. The driver's take-over process was divided into three phases, including the reaction phase, control phase, and recovery phase. Time parameters, dynamics parameters, and operation parameters were collected for each take-over phase in different obstacle avoidance scenarios. This study analyzed the variability of traffic density and take-over budget time with take-over time, lateral behavior, and longitudinal behavior. The results showed that in the reaction phase, the driver's reaction time became shorter as the scenario urgency increased. In the control phase, the steering wheel reversal rate, lateral deviation rate, braking rate, average speed, and takeover time were significantly different at different urgency levels. In the recovery phase, the average speed, accelerating rate, and take-over time differed significantly at different urgency levels. For the entire take-over process, the entire take-over time increased with the increase in urgency. The lateral take-over behavior tended to be aggressive first and then became defensive, and the longitudinal take-over behavior was defensive with the increase in urgency. The findings will provide theoretical and methodological support for the improvement of take-over behavior assistance in emergency take-over scenarios. It will also be helpful to optimize the human-machine interaction system.

Repeated conditionally automated driving on the road: How do drivers leave the loop over time?

The goal of this on the road driving study was to investigate how drivers adapt their behavior when driving with conditional vehicle automation (SAE L3) on different occasions. Specifically, we focused on changes in how fast drivers took over control from automation and how their gaze off the road changed over time. On each of three consecutive days, 21 participants drove for 50 min, in a conditionally automated vehicle (Wizard of Oz methodology), on a typical German commuting highway. Over these rides the take-over behavior and gaze behavior were analyzed. The data show that drivers’ reactions to non-critical, system initiated, take-overs took about 5.62 s and did not change within individual rides, but on average became 0.72 s faster over the three rides. After these self-paced take-over requests a final urgent take-over request was issued at the end of the third ride. In this scenario participants took over rapidly with an average of 5.28 s. This urgent take-over time was not found to be different from the self-paced take-over requests in the same ride. Regarding gaze behavior, participants’ overall longest glance off the road and the percentage of time looked off the road increased within each ride, but stayed stable over the three rides. Taken together, our results suggest that drivers regularly leave the loop by gazing off the road, but multiple exposures to take-over situations in automated driving allow drivers to come back into loop faster.

Real‐Time Non‐Driving Behavior Recognition Using Deep Learning‐Assisted Triboelectric Sensors in Conditionally Automated Driving

AbstractReal‐time recognition of non‐driving behaviors is of great importance in conditionally automated driving, as it determines the takeover time budget, which in turn has a huge impact on the performance of the takeover. Here, a novel real‐time non‐driving behavior recognition system (RNBRS) integrating self‐powered, low‐cost, easy‐to‐manufacture triboelectric sensors, and a deep learning model is proposed. The structure, working mechanism, and electrical characteristics of triboelectric sensors are investigated and analyzed. Through the ingenious structural design of single‐electrode triboelectric sensors and driving simulation experiments under conditional automated driving, non‐driving behaviors are captured in the form of electrical signals. A well‐trained long short‐term memory network model is adopted to recognize the five most typical non‐driving behaviors, including phone, console touchpad, driving, monitoring driving, and no operation, and test accuracy of 93.5% is achieved. Demonstration of a set of controlled experiments shows that RNBRS enables vehicles with conditional automation to dynamically adjust takeover time budget based on driver behavior, therefore significantly improving both safety and stability of takeover. This study opens new frontiers for the development of self‐powered electronics and inspires new thoughts on human‐machine interaction and the safety of autonomous vehicles.

Exploring the Impact of the Takeover Time for Conditionally Automated Driving Vehicles on Traffic Flow in Highway Merging Area

Human-machine handover of conditionally automated driving vehicles (CADVs) significantly affects traffic safety. Therefore, a simulation modeling was conducted for the traffic flow mixed with manual driving vehicles, fully automated driving vehicles (FADVs), and CADVs under different takeover times to unravel the impact of CADVs on traffic flow. The results showed that different takeover times significantly affected traffic flow stability. A moderate takeover time allows the driver to complete the takeover quickly with sufficient observation of the surrounding traffic conditions and mitigate the adverse effects of CADVs takeover transition on traffic flow and improve traffic flow safety. Taking moderate takeover time (7s) as the given takeover time, we developed a traffic flow model, and it is found that increasing the total penetration rates of CADVs and FADVs or that of CADVs alone will expand the traffic flow stability area. Moreover, the improving effects of traffic flow stability increase with the value of both penetration rates. This research can be a reference for the safety analysis of heterogeneous traffic flow mixed with CADVs.

Responses to Take-Over Request in Autonomous Vehicles: Effects of Environmental Conditions and Cues

Conditional autonomous vehicles have attracted significant attention, and their wide-spread use is expected to further increase in the near future. However, such vehicles can send take-over requests (TORs) to a driver who may struggle to accommodate the request while immersed in non-driving-related tasks. Previous studies have focused on TOR times and cues; however, the effects of environmental conditions have not been examined rigorously. This study, with the aim of addressing the aforementioned issue, is divided into two parts: in Study 1, we examine driver responses to TORs under different environmental conditions (i.e., sunny, rainy, snowy, foggy, and night-time), and in Study 2, we examine the effects of the proposed TOR cues (augmented reality + smartphone alert) under different environmental conditions. For this investigation, a driving simulator was used for the participants’ safety. Each study involves 33 participants. The results of Study 1 indicate significant differences in the take-over time, lane-change time, time-to-collision, maximum acceleration, and subjective mental workload corresponding to different environmental conditions. Furthermore, the results of Study 2 suggest that the proposed TOR cues significantly reduce the effects of environmental conditions on various take-over performances. We discuss the implication of these results in terms of the improvements in responses to TORs and investigation of the effects of environmental conditions on the responses to TORs.