The investigation of automated vehicle acceptance (AVA) has received considerable attention in the past few years. Understanding the factors impacting their acceptance is pivotal to ensure a large-scale and wide acceptance of AVs. The AVA by pedestrians is still little understood. To address this knowledge gap, the main objective of this study is to develop and validate an instrument for the assessment of AVA by pedestrians. We tested this instrument on a German sample of pedestrians (n = 136), considering their individual demographic characteristics, and level of affinity for technology interaction. A four-step approach was adopted to analyze the data. First, a principal component analysis was performed to reduce the number of items, exploring the sources of variation in the dataset. Second, the principal components were subjected to a confirmatory factor analysis to investigate the validity and reliability of the proposed measurement model. Third, structural equation modeling was conducted to estimate the path relationships between our constructs. The study has revealed differences between the effect sizes and significance levels of the factors influencing pedestrians’ AVA. The AVA by pedestrians was most strongly influenced by affinity for technology interaction (i.e., extent to which the individual actively approaches or avoids the interaction with new systems), performance expectancy (i.e., extent to which the individual believes that using the system will support them in achieving gains in the performance of the task) and social influence (i.e., extent to which the individual believes that people important to them think that the individual should perform the behavior). Male pedestrians were more likely to accept AVs. We also revealed significant interaction effects of age on the variables in our model. With this work, we have contributed to the development and validation of the Automated Vehicle Acceptance Questionnaire for Pedestrians (AVAQ-P). We recommend future research to replicate the study with a larger, more representative and gender-diverse population of pedestrians, considering cross-cultural differences in AVA.

Shared autonomous shuttles (SASs) could improve the mobility infrastructure in the worlds’ growing cities. This novel service could reduce congestion and improve both mobility and sustainability. To facilitate the implementation of SASs, more research is needed on the psychological aspects of sharing a small, intimate shuttle with strangers. The current study is among the first to use open-ended questions to investigate SAS acceptance. This investigation is based on the Multi-Level Model on Automated Vehicle Acceptance (MAVA). We had 236 participants answer short-form interviews including both open-ended questions and quantitative items. Quantitative data were analyzed using descriptive statistics and correlations, and qualitative data analyzed with directed content analysis. Respondents seem very positive about the proposed new transport service. We found that perceived usefulness, hedonic motivation, trust, and social influence shared large correlations with intentions to use. Other factors such as demographics, technology savviness and use of public transport did not share a linear relationship with intentions to use. Qualitative analysis suggests that, while most people do not mind sharing shuttles with strangers, some could find the social situation deterring. People seem most concerned with availability, effectiveness, travel cost and safety. The reported positive attitudes towards the service seem predicated upon trust in the government regulation and proper testing of the technology, that many think of as immature. Regulation and thorough testing may be paramount in keeping people positive. This study emphasizes the importance of trust and safety to adoption of SAS, while suggesting new factors that need further investigation.

The integration of Advanced Driver Assistance Systems (ADAS) in vehicles marks an advancement in automotive safety and driving efficiency. However, to obtain the benefits of ADAS, drivers need to understand and utilise the systems properly. This study investigates the strategies employed by U.S. drivers to learn about and operate ADAS, focusing on both the learning channels through which they acquire information and the learning content of that information. Through in-depth interviews with twenty-three drivers, who are experienced with SAE Level 2 partial driving automation, this study explores the learning methods, experiences, and knowledge drivers gain when interacting with ADAS features. The findings show that trial-and-error is the predominant method drivers use to learn about ADAS. In addition, dealerships are critical points for providing initial information, but there is variability in the quality and timing of information provided by sales personnel. In the initial learning phase, it is important to convey critical information, such as the basic functionality, operation, and limitations, while avoiding information overload. In addition, drivers expressed a desire to receive feedback on the status of these features and the reasons for any occurrences while driving. The study highlights the necessity of comprehensive learning strategies incorporating multiple learning channels, including driving schools, dealerships, digital resources, and practical experience through trial-and-error. This research shows the importance of a structured and adaptable approach to ADAS learning, tailored to drivers’ preferences and behaviours, to maximise the safety and utility of driving automation.

Driverless, SAE Level 4 automated vehicles (AVs)—vehicles operating without on-board human operators—have become operational in some cities in the U.S. The driving style and behaviors of AVs can induce changes in the behavior of road users interacting with AVs in traffic. Prior research has not collected data from road users residing in areas in which AVs are deployed and who have solid experience with AVs by regular interactions with them. As a result, a comprehensive and rich analysis of road users’ responses to AVs in traffic based on solid experience and the underlying reasons is missing. The two main research questions of this study are: 1) How do road users respond to AVs in traffic? and 2) Which factors affect road users’ responses to AVs in traffic? Semi-structured interviews were conducted with individuals residing in U.S. cities in which driverless AVs are deployed to explore how and why road users respond to driverless AVs in traffic. Content analysis was applied to manually identify themes in the data, complemented by using large language models. We also computed Spearman rank-order correlations to determine significant associations between the sub-themes. The most common road user behaviors were being more cautious around AVs, letting the AV pass and waving and gawking at them. Road users took advantage of the capabilities of AVs, cutting them off, slowing them down, or recklessly crossing the road in front. The AV safety operators typically monitored the operation of the AV, contributing to the perception that AVs are safe and predictable. Other participants reported incidences of inattentive drivers / human operators of Tesla’s SAE Level 2 partially automated driving system, being observed sleeping in the AV and rear-ending one of our participants. The most common external communication cue between road users and human drivers was eye contact, in some cases also when there was no operator present. Media reports / personal stories involving fatal accidents with AVs, particularly those linked to Tesla’s partially automated driving system, were linked to concerns about AV safety. Our study reveals significant associations between the behavior of AVs (e.g., AV being stuck) and road users’ changes in behavior, cognition (e.g., trust, distrust) and affect (e.g., perceived safety, frustration or anger). More trials with AVs on public roads can promote the interest and curiosity of road users, and their acceptance and use of AVs. The need for eHMIs and their effectiveness in promoting safer, more efficient, and comfortable interactions needs to be further investigated.

Human–Machine Interfaces (HMIs) in passenger cars have become more complex over the years, with touch screens replacing physical buttons and with layered menu-structures. This can lead to distractions. The purpose of this study is to investigate how often vehicle controls are used while driving and which underlying factors contribute to usage. Thirty drivers were observed during driving a familiar route twice, in their own car and in an unfamiliar car. In a 2 × 1 within-subject design, the experimenter drove along with each participant and used a predefined checklist to record how often participants interacted with specific functions of their vehicle while driving. The results showed that, in the familiar car, direction indicators are the most frequently used controls, followed by adjusting radio volume, moving the sun visor, adjusting temperature and changing wiper speed. Factors that influenced task frequencies included car familiarity, gender, age and weather conditions. The type of car also appears to impact task frequency. Participants interacted less with the unfamiliar car, compared to their own car, which may indicate drivers are regulating their mental load. These results are relevant for vehicle HMI designers to understand which functions should be easily and swiftly available while driving to reduce distraction by the HMI design.

Introduction: Shared automated vehicles (SAVs) could significantly enhance public transport by addressing urban mobility challenges. However, public acceptance of SAVs remains under-studied, particularly regarding how informational factors and individual personality traits influence acceptance. Methods: This study explores SAV acceptance using data from an experimental survey of 1902 respondents across Norway. Participants were randomly presented with different informational conditions about SAV services, manipulating vehicle autonomy (fully autonomous vs. steward onboard), seating orientation (facing direction of travel vs. facing other passengers), and ethnicity of co-passengers. Personality traits from the Five Factor Model (FFM) and Social Dominance Orientation (SDO) were assessed. The General Acceptance Factor (GAF), derived from the Multi-Level Model of Automated Vehicle Acceptance (MAVA), was used as the primary outcome measure. Results: No significant main or interaction effects were found from the experimentally altered information conditions. However, personality traits significantly influenced acceptance. Specifically, higher openness and agreeableness positively predicted SAV acceptance, while higher neuroticism and social dominance orientation negatively predicted acceptance. Discussion: The absence of experimental effects suggests either a limited role of the manipulated factors or insufficiently robust manipulations. Conversely, the substantial impact of personality traits highlights the importance of psychological factors, particularly trust, openness, and social attitudes, in shaping SAV acceptance. These findings emphasize the need for tailored communication strategies to enhance SAV uptake, addressing specific psychological profiles and fostering trust in automation.

Introduction: Partially automated cars are on the road. Trust in automation and perceived safety are critical factors determining use of automation. Background: Drivers misuse partially automated driving systems. Misuse is associated with mis-calibrated trust in the automation. Research gap: Little is known about the factors impacting the perceived safety when using partial driving automation. Research objective: The main objective of the present study is to provide a comprehensive driver perspective on the psychological aspects of automation use pertaining to trust in automation, perceived safety, and its relationship with use of automation. Method: Semi-structured interviews (n = 103) were conducted with users of partially automated driving systems. Supplemented with content analysis, natural language processing (NLP) techniques were applied to perform automatic text processing. Guided seed-term analysis was conducted to identify the number of occurrences of the subcategories in the dataset. Main results: We identified human operator-related, automation-related, and environmental factors of trust and perceived safety. The identified factors were more strongly associated with perceived safety than with trust. Participants with physical and visual impairments reported to feel safer using the automation compared to driving manually. Neurotic behavior during manual driving contributed to lower trust and perceived safety using the automation. A correct mental model of the capabilities and limitations of the automation did not guarantee proper automation use. A novel conceptual, process-oriented model, titled PTS-a (predicting trust in and perceived safety of automation use), synthesizes the results of the data analysis. Informed by the cognition-leads-to-emotions approach, the model posits that trust as cognition precedes perceived safety as affective construct. Trust and perceived safety determine how human operators (mis-, dis-)use the automation. Future research: We recommend future research to perform experimental studies to identify cognitive-related thoughts and beliefs pertaining to trust in automation and perceived safety to contribute to the operationalization of these constructs, and unravel the nature of their relationship.

User comfort in higher-level Automated Vehicles (AVs, SAE Level 4+) is crucial for public acceptance. AV driving styles, characterised by vehicle kinematic and proxemic factors, affect user comfort, with “human-like” driving styles expected to provide natural feelings. We investigated a) how the kinematic and proxemic factors of an AV's driving style affect users' evaluation of comfort and naturalness, and b) how the similarities between automated and users' manual driving styles affect user evaluation. Using a motion-based driving simulator, participants experienced three Level 4 automated driving styles: two human-like (defensive, aggressive) and one machine-like. They also manually drove the same route. Participants rated their comfort and naturalness of each automated controller, across twenty-four varied UK road sections. We calculated maximum absolute values of the kinematic and proxemic factors affecting the AV's driving styles in longitudinal, lateral, and vertical directions, for each road section, to characterise the automated driving styles. The Euclidean distance between AV and manual driving styles, in terms of kinematic and proxemic factors, was calculated to characterise the human-like driving style of the AV. We used mixed-effects models to examine a) the effect of AV's kinematic and proxemic factors on the evaluation of comfort and naturalness, and b) how similarities between manual and automated driving styles affected the evaluation. Results showed significant effects of lateral and rotational kinematic factors on comfort and naturalness, with longitudinal kinematic factors having a less prominent effect. Similarities in vehicle metrics, such as speed, longitudinal jerk, lateral offset, and yaw, between manual and automated driving styles, enhanced user comfort and naturalness. This research facilitates an understanding of how control features of AVs affect user experience, contributing to the design of user-centred controllers and better acceptance of higher-level AVs.

The driving style of an automated vehicle (AV) needs to be comfortable to encourage the broad acceptance and use of this newly emerging transport mode. However, current research provides limited knowledge about what influences comfort, how this concept is described, and how it is measured. This knowledge is especially lacking when comfort is linked to the AV's driving styles. This paper presents results from an online workshop with nine experts, all with hands-on experience of AVs and a long track record of research in this context. Using online tools, experts were invited to introduce concepts they considered relevant to comfort/discomfort in currently available modes of transport which offer a ride (taxi/bus/train) to users and compare these to the concepts used to define comfort and discomfort in AVs. Results showed that a wide range of terms were used to describe user comfort and discomfort for both modes. Although all terms used for existing vehicles were found to apply to AVs, additional terms were proposed for determining comfort/discomfort of AVs. For example, to enhance comfort in AVs, designers should consider good communication channels, as well as ensuring that the AV's capabilities match users’ expectations. Results also revealed that more terms were used, overall, to define discomfort, and that a comfortable ride in AVs is not just about mitigating discomfort. New concepts specific to AVs were also revealed when considering what increases their discomfort, such as whether riders’ safety and privacy are affected, or if they feel in control. Experts’ input from the workshop was used to enhance and expand a simple conceptual framework, explaining how AV driving styles, as well as other, non-driving-related factors, affect user comfort. It is hoped that this framework provides a more comprehensive list of the concepts affecting user comfort, also allowing more accurate measurement of the concept. As well as allowing for a more accurate comparison between empirical studies measuring comfort in AVs, this study will facilitate the design of more comfortable and acceptable automated driving for future vehicles.

The emerging use of automated driving systems introduces novel situations that may affect the safety of vulnerable road users such as cyclists. In this paper, we explain and conceptualise the phenomenon of phantom braking – sudden and unexpected deceleration – in automated vehicles. We apply signal detection theory to interpret phantom braking as a by-product of automated decision-making, with the vehicle favouring the avoidance of accidents at the cost of potentially causing rear-end accidents. To illustrate phantom braking and its effects on cyclists, we used a newly developed cycling simulator. An exploratory measurement conducted with a single cyclist participant revealed a possible complacency effect of the cyclist, with the cyclist’s decision-making mirroring the automated vehicle’s decision-making. The findings provide a testament to using cycling simulators for further exploration of the effects of phantom braking on cyclists.

Marie Skłodowska-Curie Actions; Innovative Training Networks (ITN); SHAPE-IT; Grant number 860410

DOI: 10.54941/ahfe1005212

Automated vehicles pose a unique challenge to the safety of vulnerable road users. Research on cyclist-automated vehicle interaction has received relatively little attention compared to pedestrian safety. This exploratory study aims to bridge this gap by identifying cyclist-automated vehicle scenarios and providing recommendations for future research. In this study, we triangulated three sources: a systematic literature review of previous research on cyclists and automated vehicles, group interviews with eight traffic safety and automation experts, and questionnaire data. The resulting scenario collection comprised 20 prototypical scenarios of cyclist-automated vehicle interaction, grouped into four categories based on the road users’ direction of movement: crossing, passing, overtaking, and merging scenarios. The survey results indicated that right-turning vehicles, dooring scenarios, and more complex situations have the highest likelihood of accidents. Passing and merging scenarios are particularly relevant for studying automated vehicle communication solutions since they involve negotiation. Future research should also consider phantom braking and driving styles of vehicles, as well as programming proactive safety behaviours and designing on-vehicle interfaces that accommodate cyclists.

Marie Skłodowska-Curie Actions; Innovative Training Networks (ITN); SHAPE-IT; Grant number 860410

Publication date: 31 December 2023

DOI: 10.1016/j.trip.2023.100986

Sharp curves in freeways are known to be unsafe design elements since drivers do not expect them. It is difficult for drivers to estimate the radius of a curve. Therefore, drivers are believed to use other cues to decelerate when approaching a curve. Based on previous successful experiences of driven speeds in curves, drivers are thought to have built expectations of safe speeds given certain cues, minimalising risks. This research employs a Bayesian Belief Network to model driver expectations using measured speeds in 153 curves and data on the characteristics of the curve approaches. This model mimics expectations as the probability of measured speeds given certain cues. Using Bayes theorem, prior beliefs on safe speeds are updated towards a posterior belief when a new cue is observed during curve approach. We refer to this posterior belief as expected safe speed. Drivers are assumed to adjust their operating speed if it does not match their expected safe speed. The model shows that the visible deflection angle has a large influence in setting the expectations of a safe speed for an upcoming curve. In addition, the preceding type of roadway and the number of lanes are both important cues to set a driver's expectations of a safe speed. Speed and warning signs are shown to be interdependent on the road scene and hence have less influence in setting expectations. This research shows that design and safety assessment of freeway curves should be considered aligned with the road scene upstream of the curve.

The primary aim of this study was to develop an accurate measure of acceptance for shared autonomous vehicles (SAVs) and to assess whether this measure can predict intentions to use SAVs. One leading model for explaining technology uptake is the UTAUT (Unified theory of acceptance and use of technology). This model is extensive and has received numerous suggested extensions and revisions, even being developed into a Multi-Level Model of Autonomous Vehicle Acceptance (MAVA). The challenge is to consolidate a model that effectively measures SAV acceptance and to determine which extensions capture the unique social situation within SAVs. The current study used survey data from 1902 respondents. The sample was split into two: one half underwent a principal component analysis (PCA) and the other half a confirmatory factor analysis (CFA). We found that the 24 items we included were reducible to a single general acceptance factor (GAF), with three additional factors measuring interpersonal security, sociability, and attractivity. The GAF was, by a large margin, the most efficacious predictor of intention to use SAVs. The GAF could be further reduced to as little as two predictors, trust and usefulness, accounting for over 70 % of the variance in intention to use. However, there is also an argument to be made that the other components of SAV acceptance may capture different nuances of the service, particularly relating to the social situation. Interaction terms show differences between genders in their rating of sociability and how this impacts intentions to use SAVs. Our findings carry significant implications for future research in this field. They underscore the pivotal roles of trust and usefulness while corroborating the notion that SAV acceptance is best represented by a single latent component. However, further investigation is warranted to explore individual-level moderating effects on the other components, potentially offering novel insights for the design of future SAV services.

This study investigates acceptance of shared autonomous shuttles (SASs) in a suburban area. A model where contextual variables were mediated through trust in SASs and technology optimism was tested. We examined intentions to use SASs without a steward and the significance of social distancing. Data were collected at the start and end of a 2020–2021 pilot involving 922 and 608 participants respectively, operating at SAE level 3. Findings indicate that trust and technological optimism significantly influence the willingness to use SASs, though contextual variables show minimal impact. Older adults and women displayed lower trust and optimism, reducing their usage intentions. These two groups also feel that it is more important to be able to keep social distance while riding SASs. The study suggests that future pilots should avoid negative impacts from using immature technology and address the social needs of specific groups.

Interaction with vulnerable road users in complex urban traffic environments poses a significant challenge for automated vehicles. Solutions to facilitate safe and acceptable interactions in future automated traffic include equipping automated vehicles and vulnerable road users, such as cyclists, with awareness or notification systems, as well as connecting road users to a network of motorised vehicles and infrastructure. This paper provides a synthesis of the current literature on communication technologies, systems, and devices available to cyclists, including technologies present in the environment and on motorised interaction partners such as vehicles, and discusses the outlook for technology-driven solutions in future automated traffic. The objective is to identify, classify, and count the technologies, systems, and devices that have the potential to aid cyclists in traffic with automated vehicles. Additionally, this study aims to extrapolate the potential benefits of these systems and stimulate discourse on the implications of connected vulnerable road users. We analysed and coded 92 support systems using a taxonomy of 13 variables based on the physical, communicational, and functional attributes of the systems. The discussion frames these systems into four categories: cyclist wearables, on-bike devices, vehicle systems, and infrastructural systems, and highlights the implications of the visual, auditory, motion-based, and wireless modes of communication of the devices. The most common system was cyclist wearables (39%), closely followed by on-bike devices (38%) and vehicle systems (33%). Most systems communicated visually (77%). We suggest that interfaces on motorised vehicles accommodate cyclists with visibility all around the car and incorporate two-way communication. The type of system and the effect of communication modality on performance and safety needs further research, preferably in complex and representative test scenarios with automated vehicles. Finally, our study highlights the ethical implications of connected road users and suggests that the future outlook of transport systems may benefit from a more inclusive and less car-centred approach, shifting the burden of safety away from vulnerable road users and promoting more cyclist-friendly solutions.

Marie Skłodowska-Curie Actions; Innovative Training Networks (ITN); SHAPE-IT; Grant number 860410

Publication date: 6 May 2023

DOI: 10.1016/j.apergo.2023.104043

In future traffic, intelligent user interfaces may aid cyclists in interpreting the behaviour of automated vehicles. Cyclists can be equipped with obstacle-detecting sensors, and an interface could display relevant information or use audible alerts to warn or inform cyclists of other road users' intent and potential hazards. Researching intelligent user interfaces for cyclists is vital for understanding how users can efficiently and safely interact with automated vehicles. This work-in-progress paper presents two studies for developing and testing user interfaces for cyclists in future automated traffic. In the first study, we reanalysed interview data from 30 cyclists, resulting in two interface concepts: the app CycleSafe and an omnidirectional on-vehicle interface capable of communicating cyclist recognition. In the second study, we outline an envisioned experiment to test these two concepts in a naturalistic environment with cyclists and a vehicle emulating automation. We hypothesise that cyclists prefer receiving warning signals over no warnings, prefer early over late warnings, and that auditory signals and visual on-vehicle interfaces will perform better than visual on-bike interfaces.

Marie Skłodowska-Curie Actions; Innovative Training Networks (ITN); SHAPE-IT; Grant number 860410

Publication date: 27 March 2023

DOI: 10.1145/3581754.3584140

Tesla's Full Self-Driving Beta (FSD) program introduces technology that extends the operational design domain of standard Autopilot from highways to urban roads. This research conducted 103 in-depth semi-structured interviews with users of Tesla's FSD Beta and standard Autopilot to evaluate the impact on user behavior and perception. It was found that drivers became complacent over time with Autopilot engaged, failing to monitor the system, and engaging in safety-critical behaviors, such as hands-free driving, enabled by weights placed on the steering wheel, mind wandering, or sleeping behind the wheel. Drivers' movement of eyes, hands, and feet became more relaxed with experience with Autopilot engaged. FSD Beta required constant supervision as unfinished technology, which increased driver stress and mental and physical workload as drivers had to be constantly prepared for unsafe system behavior (doing the wrong thing at the worst time). The hands-on wheel check was not considered as being necessarily effective in driver monitoring and guaranteeing safe use. Drivers adapt to automation over time, engaging in potentially dangerous behaviors. Some behavior seems to be a knowing violation of intended use (e.g., weighting the steering wheel), and other behavior reflects a misunderstanding or lack of experience (e.g., using Autopilot on roads not designed for). As unfinished Beta technology, FSD Beta can introduce new forms of stress and can be inherently unsafe. We recommend future research to investigate to what extent these behavioral changes affect accident risk and can be alleviated through driver state monitoring and assistance.

Sufficient cycle track width is important to prevent single-bicycle crashes and collisions between cyclists. The assumptions on which the minimum width is based in guidelines is founded on only a few studies. The aim of the present study is to investigate the relationship between cycle track width and lateral position of cyclists. Method: We conducted an experiment to evaluate the lateral position of cyclists along cycle tracks with different widths (Study 1). Participants cycled on an instrumented bicycle with a LIDAR to measure their lateral position. Five conditions were defined: cycle track width of 100 cm, 150 cm and 200 cm without interaction, and cycle track width of 150 cm and 200 cm with an oncoming cyclist simulated by a parked bicycle. The cross-sectional Study 2 is based on the collected lateral position measurements at cycle tracks with varying width reported in Dutch studies since 2010. Results: The experimental Study 1 with 24 participants shows that an increase in cycle track width causes cyclists to ride further away from the verge and keep more distance from an oncoming cyclist. The cross-sectional Study 2 was based on lateral position measured at 33 real-life Dutch cycle tracks. Study 2 yielded similar results, indicating that doubling pavement width increases lateral position by some 50%. Study 2 shows that, compared with a solo cyclist without interaction, a right-hand cyclist of a duo and a cyclist meeting an oncoming cyclist ride around 30% closer to the verge. Conclusions: The wider the cycle track, the more distance cyclists maintain from the verge. Cyclists ride closer to the verge due to oncoming cyclists. Practical applications: Given a cyclists’ lateral position while meeting, common variations between cyclists’ steering behavior, and vehicle width and circumstances, a cycle track width of 250 cm is needed for safe meeting maneuvers.

Although much research is done on speed and gaze behaviour inside curves, there is little understanding of which cues drivers use to anticipate and slow down while approaching curves. Therefore, an on road experiment was conducted in which 31 participants drove through six freeway curves in their own car. During the experiment, look-ahead fixations and speed were recorded using an eye-tracker and a GPS tracker, respectively. In addition to these measurements, the participants verbalised their reasons for changing speed. The distribution of fixations over various areas of interest was investigated around the start of deceleration before each curve and around the start of each curve. Verbalisation data were analysed to infer the number and types of reasons for changing speed and when these were mentioned together with mentions of deceleration before a curve. The results showed that before starting to decelerate, the participants fixated mostly on the Focus of Expansion and edges parallel to the curve trajectory, whereas most fixations on warning or speed signs were recorded mostly after participants started to decelerate. These findings suggest that drivers use information from the Focus of Expansion, be it a change in optical flow or the presence of a kink in the alignment, as the main cue to start decelerating. Parallel edges are also important cues, whereas warning and speed signs are primarily used to confirm that a speed change is needed.