Automated driving systems (ADSs) promise a safe, comfortable and efficient driving experience. However, fatalities involving vehicles equipped with ADSs are on the rise. The full potential of ADSs cannot be realized unless the robustness of state-of-the-art is improved further. This paper discusses unsolved problems and surveys the technical aspect of automated driving. Studies regarding present challenges, high-level system architectures, emerging methodologies and core functions including localization, mapping, perception, planning, and human machine interfaces, were thoroughly reviewed. Furthermore, many state-of-the-art algorithms were implemented and compared on our own platform in a real-world driving setting. The paper concludes with an overview of available datasets and tools for ADS development.

A Survey of Autonomous Driving: Common Practices and Emerging Technologies

This study investigated user acceptance, worries, and willingness to buy partially, highly, and fully automated vehicles. By means of a 63-question Internet-based survey, we collected 5,000 responses from 109 countries (40 countries with at least 25 respondents). We determined cross-national differences, and assessed correlations with personal variables, such as age, gender, and personality traits as measured with a short version of the Big Five Inventory. Results showed that respondents, on average, found manual driving the most enjoyable mode of driving. Responses were diverse: 22% of the respondents did not want to pay more than $0 for a fully automated driving system, whereas 5% indicated they would be willing to pay more than $30,000 for it, and 33% indicated that fully automated driving would be highly enjoyable. 69% of respondents estimated that fully automated driving will reach a 50% market share between now and 2050. Respondents were found to be most concerned about software hacking/misuse, and were also concerned about legal issues and safety. More neurotic people were slightly less comfortable about data transmitting, whereas agreeable persons were slightly more comfortable with that. The more developed countries (in terms of accident statistics, education, and income) were less comfortable with their vehicle transmitting data, with cross-national correlations between ρ= -0.80 and ρ= -0.90. The present results indicate the major areas of promise and concern among the international public, and could be useful for vehicle developers and other stakeholders.

/pdf/public-opinion-on-automated-driving-results-of-an-49ui4752zp.pdf

Public Opinion on Automated Driving: Results of an International Questionnaire Among 5,000 Respondents

This study investigated user acceptance, concerns, and willingness to buy partially, highly, and fully automated vehicles. By means of a 63-question Internet-based survey, we collected 5000 responses from 109 countries (40 countries with at least 25 respondents). We determined cross-national differences, and assessed correlations with personal variables, such as age, gender, and personality traits as measured with a short version of the Big Five Inventory. Results showed that respondents, on average, found manual driving the most enjoyable mode of driving. Responses were diverse: 22% of the respondents did not want to pay more than $0 for a fully automated driving system, whereas 5% indicated they would be willing to pay more than $30,000, and 33% indicated that fully automated driving would be highly enjoyable. 69% of respondents estimated that fully automated driving will reach a 50% market share between now and 2050. Respondents were found to be most concerned about software hacking/misuse, and were also concerned about legal issues and safety. Respondents scoring higher on neuroticism were slightly less comfortable about data transmitting, whereas respondents scoring higher on agreeableness were slightly more comfortable with this. Respondents from more developed countries (in terms of lower accident statistics, higher education, and higher income) were less comfortable with their vehicle transmitting data, with cross-national correlations between ρ = −0.80 and ρ = −0.90. The present results indicate the major areas of promise and concern among the international public, and could be useful for vehicle developers and other stakeholders.

/pdf/public-opinion-on-automated-driving-results-of-an-4z5lglcxwz.pdf

Public opinion on automated driving: results of an international questionnaire among 5000 respondents

Adaptive cruise control (ACC), a driver assistance system that controls longitudinal motion, has been introduced in consumer cars in 1995. A next milestone is highly automated driving (HAD), a system that automates both longitudinal and lateral motion. We investigated the effects of ACC and HAD on drivers' workload and situation awareness through a meta-analysis and narrative review of simulator and on-road studies. Based on a total of 32 studies, the unweighted mean self-reported workload was 43.5% for manual driving, 38.6% for ACC driving, and 22.7% for HAD (0% = minimum, 100 = maximum on the NASA Task Load Index or Rating Scale Mental Effort). Based on 12 studies, the number of tasks completed on an in-vehicle display relative to manual driving (100%) was 112% for ACC and 261% for HAD. Drivers of a highly automated car, and to a lesser extent ACC drivers, are likely to pick up tasks that are unrelated to driving. Both ACC and HAD can result in improved situation awareness compared to manual driving if drivers are motivated or instructed to detect objects in the environment. However, if drivers are engaged in non-driving tasks, situation awareness deteriorates for ACC and HAD compared to manual driving. The results of this review are consistent with the hypothesis that, from a Human Factors perspective, HAD is markedly different from ACC driving, because the driver of a highly automated car has the possibility, for better or worse, to divert attention to secondary tasks, whereas an ACC driver still has to attend to the roadway.

/pdf/effects-of-adaptive-cruise-control-and-highly-automated-zm9ko1q3vg.pdf

Effects of adaptive cruise control and highly automated driving on workload and situation awareness: A review of the empirical evidence

A driving simulator study was designed to investigate drivers’ ability to resume control from a highly automated vehicle in two conditions: (i) when automation was switched off and manual control was required at a system-based, regular interval and (ii) when transition to manual was based on the length of time drivers were looking away from the road ahead. In addition to studying the time it took drivers to successfully resume control from the automated system, eye tracking data were used to observe visual attention to the surrounding environment and the pattern of drivers’ eye fixations as manual control was resumed in the two conditions. Results showed that drivers’ pattern of eye movement fixations remained variable for some time after automation was switched off, if disengagement was actually based on drivers’ distractions away from the road ahead. When disengagement was more predictable and system-based, drivers’ attention towards the road centre was higher and more stable. Following a lag of around 10 s, drivers’ lateral control of driving and steering corrections (as measured by SDLP and high frequency component of steering, respectively) were more stable when transition to manual control was predictable and based on a fixed time. Whether automation transition to manual was based on a fixed or variable interval, it took drivers around 35–40 s to stabilise their lateral control of the vehicle. The results of this study indicate that if drivers are out of the loop due to control of the vehicle in a limited self-driving situation (Level 3 automation), their ability to regain control of the vehicle is better if they are expecting automation to be switched off. As regular disengagement of automation is not a particularly practical method for keeping drivers in the loop, future research should consider how to best inform drivers of their obligation to resume control of driving from an automated system.

Transition to manual: driver behaviour when resuming control from a highly automated vehicle

Precis: This study examined the effect of changes in workload on performance in highly automated and manual driving. Variations in workload were also observed using blink measures. Results showed good driver response to incidents in the highly automated condition and some predictions in workload levels by blink frequency measures.

/pdf/highly-automated-driving-secondary-task-performance-and-vdxkls7664.pdf

Highly automated driving, secondary task performance, and driver state.

Previous research has indicated that high levels of vehicle automation can result in reduced driver situation awareness, but has also highlighted potential benefits of such future vehicle designs through enhanced safety and reduced driver workload. Well-designed automation allows drivers’ visual attention to be focused away from the roadway and toward secondary, in-vehicle tasks. Such tasks may be pleasant distractions from the monotony of system monitoring. This study was undertaken to investigate the impact of voluntary secondary task uptake on the system supervisory responsibilities of drivers experiencing highly-automated vehicle control. Independent factors of Automation Level (manual control, highly-automated) and Traffic Density (light, heavy) were manipulated in a repeated-measures experimental design. 49 drivers participated using a high-fidelity driving simulator that allowed drivers to see, hear and, crucially, feel the impact of their automated vehicle handling. Drivers experiencing automation tended to refrain from behaviours that required them to temporarily retake manual control, such as overtaking, resulting in an increased journey time. Automation improved safety margins in car following, however this was restricted to conditions of light surrounding traffic. Participants did indeed become more heavily involved with the in-vehicle entertainment tasks than they were in manual driving, affording less visual attention to the road ahead. This might suggest that drivers are happy to forgo their supervisory responsibilities in preference of a more entertaining highly-automated drive. However, they did demonstrate additional attention to the roadway in heavy traffic, implying that these responsibilities are taken more seriously as the supervisory demand of vehicle automation increases. These results may dampen some concerns over driver underload with vehicle automation, assuming vehicle manufacturers embrace the need for positive system feedback and drivers also fully appreciate their supervisory obligations in such future vehicle designs.

/pdf/behavioural-changes-in-drivers-experiencing-highly-automated-3y13za1dpz.pdf

Behavioural changes in drivers experiencing highly-automated vehicle control in varying traffic conditions

OBJECTIVE: The study was designed to show how driver attention to the road scene and engagement of a choice of secondary tasks are affected by the level of automation provided to assist or take over the basic task of vehicle control. It was also designed to investigate the difference between support in longitudinal control and support in lateral control. BACKGROUND: There is comparatively little literature on the implications of automation for drivers' engagement in the driving task and for their willingness to engage in non-driving-related activities. METHOD: A study was carried out on a high-level driving simulator in which drivers experienced three levels of automation: manual driving, semiautomated driving with either longitudinal or lateral control provided, and highly automated driving with both longitudinal and lateral control provided. Drivers were free to pay attention to the roadway and traffic or to engage in a range of entertainment and grooming tasks. RESULTS: Engagement in the nondriving tasks increased from manual to semiautomated driving and increased further with highly automated driving. There were substantial differences in attention to the road and traffic between the two types of semiautomated driving. CONCLUSION: The literature on automation and the various task analyses of driving do not currently help to explain the effects that were found. Lateral support and longitudinal support may be the same in terms of levels of automation but appear to be regarded rather differently by drivers. Keywords: Driver distraction; Language: en

/pdf/control-task-substitution-in-semiautomated-driving-does-it-1n4cjxe5cv.pdf

Control Task Substitution in Semiautomated Driving Does It Matter What Aspects Are Automated

Forward collision warning (FCW) systems can reduce rear-end vehicle collisions. However, if the presentation of warnings is perceived as mistimed, trust in the system is diminished and drivers become less likely to respond appropriately. In this driving simulator investigation, 45 drivers experienced two FCW systems: a non-adaptive and an adaptive FCW that adjusted the timing of its alarms according to each individual driver’s reaction time. Whilst all drivers benefited in terms of improved safety from both FCW systems, non-aggressive drivers (low sensation seeking, long followers) did not display a preference to the adaptive FCW over its non-adaptive equivalent. Furthermore, there was little evidence to suggest that the non-aggressive drivers’ performance differed with either system. Benefits of the adaptive system were demonstrated for aggressive drivers (high sensation seeking, short followers). Even though both systems reduced their likelihood of a crash to a similar extent, the aggressive drivers rated each FCW more poorly than their non-aggressive contemporaries. However, this group, with their greater risk of involvement in rear-end collisions, reported a preference for the adaptive system as they found it less irritating and stress-inducing. Achieving greater acceptance and hence likely use of a real system is fundamental to good quality FCW design.

/pdf/potential-benefits-of-an-adaptive-forward-collision-warning-ja9byrrkg1.pdf

Frank Lai

Papers

Transition to manual: driver behaviour when resuming control from a highly automated vehicle

Highly automated driving, secondary task performance, and driver state.

Behavioural changes in drivers experiencing highly-automated vehicle control in varying traffic conditions

Control Task Substitution in Semiautomated Driving Does It Matter What Aspects Are Automated

Potential benefits of an adaptive forward collision warning system