TASCC: Human Interaction: Designing Autonomy in Vehicles (HI:DAV)
Lead Research Organisation:
University of Southampton
Department Name: Faculty of Engineering & the Environment
Abstract
Cars that can drive themselves have been predicted for some time, but they are nearly with us. Highly automated vehicles are likely to be on public roads within the next ten years. The largest gap in our understanding of vehicle automation is how drivers will react to this new technology and how best to design the driver-automation interaction. Our project will answer these questions by studying a wide range of drivers with different driving experience in simulators, or test-tracks and in road going vehicles.
We will start by modeling driver behaviour in our laboratories in order to help us design inclusive, user-centered, interfaces with vehicle automation. We plan to learn from different situations where automation is already used (such as the Boeing versus Airbus design approach of 'soft' and 'hard' automation). Then we will test the designs out in our driving simulator (which comprises a Jaguar XJ connected to computers with large projectors and screens). We will test drivers of different ages, gender, experience and capabilities, in a range of scenarios (e.g., different road types and environmental conditions) with different automation systems (e.g., autonomous driving, auto 'valet' parking, adaptive vehicle personalization, off-road assistance) and different interfaces (which we will design).
Our design approach will personalise the driver interfaces to the widest range of drivers possible (e.g., wide ranges of age and driving experience). We want the driver-automation interfaces to be intuitive and the necessary responses to be obvious. We design so that the system adapts to you, not you having to adapt to the system. We will also monitor driver behaviour over successive weeks of use. This will help us to understand how drivers learn to respond to the automation over time. We will be studying how control of the vehicle is handed back and forth between the automation and the driver. We are also interested to know what the driver is doing when the vehicle is under automated control. To help in this understanding, we will be monitoring the driver physiological and psychological states. The idea is to use this information to adapt the automation to the driver and the situation, so that performance of the system is optimised. This will enhance safety to the benefit of the driver and other road users.
The studies will progress from the simulator to the test-track, as our interaction and interface designs evolve with testing. On the test-track we can record driver behaviour physiological and psychological states to see what further changes are needed and whether the automation can be even more highly tailored to the situation. As the research progresses we will take the revised designs into road going vehicles for the final set of tests. These tests will be used to validate the designs and prepare them for delivery into production vehicles. At the end of the research, we will be able to provide JLR with the design methods and well as the designs themselves. This means that they will be able use the methods we have provided to design new systems into the future.
During the course of the research the Universities of Cambridge and Southampton will be working closely with JLR engineers to ensure that the UK remains at the forefront of technological innovation in vehicle automation. We will have answered the questions about how drivers will react to this new technology and how best to design the driver-automation interaction. The success of vehicle automation design will be on designing appropriate interactions and interfaces that support the driver. Our research will be essential to that success.
We will start by modeling driver behaviour in our laboratories in order to help us design inclusive, user-centered, interfaces with vehicle automation. We plan to learn from different situations where automation is already used (such as the Boeing versus Airbus design approach of 'soft' and 'hard' automation). Then we will test the designs out in our driving simulator (which comprises a Jaguar XJ connected to computers with large projectors and screens). We will test drivers of different ages, gender, experience and capabilities, in a range of scenarios (e.g., different road types and environmental conditions) with different automation systems (e.g., autonomous driving, auto 'valet' parking, adaptive vehicle personalization, off-road assistance) and different interfaces (which we will design).
Our design approach will personalise the driver interfaces to the widest range of drivers possible (e.g., wide ranges of age and driving experience). We want the driver-automation interfaces to be intuitive and the necessary responses to be obvious. We design so that the system adapts to you, not you having to adapt to the system. We will also monitor driver behaviour over successive weeks of use. This will help us to understand how drivers learn to respond to the automation over time. We will be studying how control of the vehicle is handed back and forth between the automation and the driver. We are also interested to know what the driver is doing when the vehicle is under automated control. To help in this understanding, we will be monitoring the driver physiological and psychological states. The idea is to use this information to adapt the automation to the driver and the situation, so that performance of the system is optimised. This will enhance safety to the benefit of the driver and other road users.
The studies will progress from the simulator to the test-track, as our interaction and interface designs evolve with testing. On the test-track we can record driver behaviour physiological and psychological states to see what further changes are needed and whether the automation can be even more highly tailored to the situation. As the research progresses we will take the revised designs into road going vehicles for the final set of tests. These tests will be used to validate the designs and prepare them for delivery into production vehicles. At the end of the research, we will be able to provide JLR with the design methods and well as the designs themselves. This means that they will be able use the methods we have provided to design new systems into the future.
During the course of the research the Universities of Cambridge and Southampton will be working closely with JLR engineers to ensure that the UK remains at the forefront of technological innovation in vehicle automation. We will have answered the questions about how drivers will react to this new technology and how best to design the driver-automation interaction. The success of vehicle automation design will be on designing appropriate interactions and interfaces that support the driver. Our research will be essential to that success.
Planned Impact
Jaguar Land Rover
JLR will benefit in terms of the interfaces and interaction designs that can be placed into production vehicles. We will work closely with JLR to ensure that the cutting-edge insights into designing autonomy together with the associated methods and interfaces are shared as soon as they are available. Secondment opportunities at the Universities of Cambridge and Southampton will be offered throughout the project so that they can work alongside the researchers on the project. Workshops will be held with JLR engineers to train them in the methods. A project website will be hosted. It will contain case studies, methods, and research findings for JLR and the research teams to use. At the end of the project JLR will be equipped with design guidance, design methods, final prototype adaptive interfaces and experience of working with world-leading Human Factors Engineering researchers. This will equip JLR with all the necessary expertise to help them move towards a fully autonomous car.
Vehicle Manufacturing Supply Chain
UK economic growth depends on maintaining a world-leading position in automotive research and development. Features, such as vehicle automation, are vital in this respect. The addressable market size would initially be on JLR products - approximately 400,000+ cars per year, with the majority taking autonomous driving technologies. The global car market is much bigger and, with autonomous driving systems in cars growing significantly, this technology could become a defining basis of the autonomous driving interface.
Automotive Standards Organisations
The findings from the research will inform Automotive Standards Organisations about the limits of driver performance which need to be observed in design. Such limits are likely to include the nature of the handover tasks from driver to vehicle and back again, the timeliness of driver interventions, driver state monitoring and the degree of driver engagement/disengagement with different types of automation. The research would also be able to deliver templates for driver-automation interaction and interface strategies together with recommended methods and guidelines for vehicle automation.
Wider Society
The inclusive design focus will assist in driving mobility for all. This will help to support and assist the mobility of older drivers and help in maintaining their independence and quality of life. Congestion can be reduced with automated vehicles by re-routing to avoid busy roadways and smoother driving to avoid the 'snaking' that can occur with manually operated vehicles. Automated vehicles can help to the driver to reduce emissions and achieve better fuel economy through smoother acceleration and deceleration profiles. Automation can also help to reduce driver stress by smoothing out the peaks and troughs of driving workload. Finally, an accident reduction is anticipated through automated driving supporting driver situation awareness in both longitudinal and lateral control, as well as taking-over if the driver fails to cope in an emergency. The automotive specific findings could be generalisable to other field such as aviation, nuclear and rail.
JLR will benefit in terms of the interfaces and interaction designs that can be placed into production vehicles. We will work closely with JLR to ensure that the cutting-edge insights into designing autonomy together with the associated methods and interfaces are shared as soon as they are available. Secondment opportunities at the Universities of Cambridge and Southampton will be offered throughout the project so that they can work alongside the researchers on the project. Workshops will be held with JLR engineers to train them in the methods. A project website will be hosted. It will contain case studies, methods, and research findings for JLR and the research teams to use. At the end of the project JLR will be equipped with design guidance, design methods, final prototype adaptive interfaces and experience of working with world-leading Human Factors Engineering researchers. This will equip JLR with all the necessary expertise to help them move towards a fully autonomous car.
Vehicle Manufacturing Supply Chain
UK economic growth depends on maintaining a world-leading position in automotive research and development. Features, such as vehicle automation, are vital in this respect. The addressable market size would initially be on JLR products - approximately 400,000+ cars per year, with the majority taking autonomous driving technologies. The global car market is much bigger and, with autonomous driving systems in cars growing significantly, this technology could become a defining basis of the autonomous driving interface.
Automotive Standards Organisations
The findings from the research will inform Automotive Standards Organisations about the limits of driver performance which need to be observed in design. Such limits are likely to include the nature of the handover tasks from driver to vehicle and back again, the timeliness of driver interventions, driver state monitoring and the degree of driver engagement/disengagement with different types of automation. The research would also be able to deliver templates for driver-automation interaction and interface strategies together with recommended methods and guidelines for vehicle automation.
Wider Society
The inclusive design focus will assist in driving mobility for all. This will help to support and assist the mobility of older drivers and help in maintaining their independence and quality of life. Congestion can be reduced with automated vehicles by re-routing to avoid busy roadways and smoother driving to avoid the 'snaking' that can occur with manually operated vehicles. Automated vehicles can help to the driver to reduce emissions and achieve better fuel economy through smoother acceleration and deceleration profiles. Automation can also help to reduce driver stress by smoothing out the peaks and troughs of driving workload. Finally, an accident reduction is anticipated through automated driving supporting driver situation awareness in both longitudinal and lateral control, as well as taking-over if the driver fails to cope in an emergency. The automotive specific findings could be generalisable to other field such as aviation, nuclear and rail.
Publications
Walker Guy H.
(2018)
Vehicle Feedback and Driver Situation Awareness
Stanton NA
(2017)
State-of-science: situation awareness in individuals, teams and systems.
in Ergonomics
Stanton N
(2019)
Models and methods for collision analysis: A comparison study based on the Uber collision with a pedestrian
in Safety Science
Stanton N
(2021)
Validating Operator Event Sequence Diagrams: The case of an automated vehicle to human driver handovers
in Human Factors and Ergonomics in Manufacturing & Service Industries
Stanton N
(2021)
Modelling Automation-Human Driver Handovers Using Operator Event Sequence Diagrams
in Future Transportation
Stanton N
(2019)
Thematic issue: driving automation and autonomy
in Theoretical Issues in Ergonomics Science
Stanton N
(2021)
OESDs in an on-road study of semi-automated vehicle to human driver handovers
in Cognition, Technology & Work
Stanton N
(2020)
Turing in the driver's seat: Can people distinguish between automated and manually driven vehicles?
in Human Factors and Ergonomics in Manufacturing & Service Industries
Salmon P
(2020)
Methodological issues in systems Human Factors and Ergonomics: Perspectives on the research-practice gap, reliability and validity, and prediction
in Human Factors and Ergonomics in Manufacturing & Service Industries
| Description | About the Project: This project has developed a new approach for the design of handovers between vehicle automation and the human driver. The approach brings together a range of methods from user-centred design, inclusive design and ecological interface design. This approach has been used to design customisable human-machine interfaces for automation-driver handovers and has been formally tested in driving simulators and in a Jaguar IPACE on a British motorway. Validation studies were undertaken and design guidelines were produced. This project discovered that there were links between customization and user performance. In particular, the handover times were quicker (on average, but with a greater range) with customized settings than the defaults. Drivers much preferred the customized settings to the defaults and there was no adverse affect on post-handover driving performance (in terms of lane and speed stability). This is one of the first studies of its type in the UK and Europe, with a genuine automated vehicle driving (rather than being driven by a surrogate driver from the passenger seat or rear of the vehicle). Benefits to the academia: New insights into the customisation of handovers between vehicle automation and the human driver have been generated. A new method for design of driver-vehicle interfaces has been developed, called User-Centred Ecological Interface Design (UCEID). Modelling of the handovers have been validated by comparing the predictions made in Operator Event Sequence Diagrams with videos of the behaviour of human drivers, both in simulators and on the road. An on-road comparison study (Tesla, Mercedes, IPACE) has already been published in the peer-review journal Applied Ergonomics. The remaining research has been published in a series of peer-reviewed journal articles and a book in the Transportation Human Factors series by CRC Press, entitled: Designing Interaction and Interfaces for Automated Vehicles: User-Centred Ecological Interface Design and Testing. Benefit impact/industrialisation impact to JLR: This project has uncovered a number of important factors associated with the development of automated vehicles. A number of key design requirements that JLR can take forward into their development process have been established. These include operator event sequence diagrams that describe the process of handover. The project has also developed intuitive Human Machine Interface proposals and evaluations that could form part of future vehicle interfaces. There are also novel findings around the topic of customisation that will potentially increase customer satisfaction with handovers in certain automated vehicle scenarios. This project has also created specific Intellectual Property for JLR that could be used in future vehicles related to the handover from automation to manual driving. Additionally, a process that goes from basic scientific foundation through to finalized, and evaluated, concept following a series of design iterations, called the UCEID process, has been defined and shared with JLR for use in future vehicle program design. Guidelines for the design of handovers between vehicle automation and the human driver have also been presented in the final report to JLR. |
| Exploitation Route | Vehicle manufactures can use the approaches to design the interactions and interface between humans and automation. |
| Sectors | Transport |
| Description | This project has developed a new approach for the design of handovers between vehicle automation and the human driver. The approach brings together a range of methods from user-centred design, inclusive design and ecological interface design. This approach has been used to design customisable human-machine interfaces for automation-driver handovers and has been formally tested in driving simulators and in a Jaguar IPACE on a British motorway. Validation studies were undertaken and design guidelines were produced. This project discovered that there were links between customization and user performance. In particular, the handover times were quicker (on average, but with a greater range) with customized settings than the defaults. Drivers much preferred the customized settings to the defaults and there was no adverse affect on post-handover driving performance (in terms of lane and speed stability). This is one of the first studies of its type in the UK and Europe, with a genuine automated vehicle driving (rather than being driven by a surrogate driver from the passenger seat or rear of the vehicle). Benefits to the academia: New insights into the customisation of handovers between vehicle automation and the human driver have been generated. A new method for design of driver-vehicle interfaces has been developed, called User-Centred Ecological Interface Design (UCEID). Modelling of the handovers have been validated by comparing the predictions made in Operator Event Sequence Diagrams with videos of the behaviour of human drivers, both in simulators and on the road. An on-road comparison study (Tesla, Mercedes, IPACE) has already been published in the peer-review journal Applied Ergonomics. The remaining research is being published in a series of peer-reviewed journal articles and a book in the Transportation Human Factors series by CRC Press, entitled: Designing Interaction and Interfaces for Automated Vehicles: User-Centred Ecological Interface Design and Testing. Benefit impact/industrialisation impact to JLR: This project has uncovered a number of important factors associated with the development of automated vehicles. A number of key design requirements that JLR can take forward into their development process have been established. These include operator event sequence diagrams that describe the process of handover. The project has also developed intuitive Human Machine Interface proposals and evaluations that could form part of future vehicle interfaces. There are also novel findings around the topic of customisation that will potentially increase customer satisfaction with handovers in certain automated vehicle scenarios. This project has also created specific Intellectual Property for JLR that could be used in future vehicles related to the handover from automation to manual driving. Additionally, a process that goes from basic scientific foundation through to finalized, and evaluated, concept following a series of design iterations, called the UCEID process, has been defined and shared with JLR for use in future vehicle program design. Guidelines for the design of handovers between vehicle automation and the human driver have also been presented in the final report to JLR. |
| First Year Of Impact | 2016 |
| Sector | Transport |
| Impact Types | Economic |
| Description | Jaguar Land Rover |
| Organisation | Jaguar Land Rover Automotive PLC |
| Department | Jaguar Land Rover |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Collaborating with Jaguar Land Rover on the development of new driver interfaces for vehicle automation. We have already tested competitor products in Tesla and Mercedes. |
| Collaborator Contribution | We have already tested competitor products in Tesla and Mercedes. |
| Impact | Reports to JLR on project progress |
| Start Year | 2016 |