Researchers study human reaction to sharing control with driverless car

The first trials of the Venturer autonomous vehicle project, a Government backed research project, have focussed on a critical aspect in the field of autonomous vehicle (AV) technology.

The ‘handover’, when the control of the car switches from human driver to autonomous mode, is one of the least studied aspects of this developing technology. However, the interaction between the human driver and the car is a vital component in taking the new technology forward to a stage where it can be safely and successfully deployed on public roads.

The first trials tested drivers in a static simulator and in the Venturer driverless car, on private roads at the University of the West of England (UWE Bristol). Under controlled conditions at speeds up to 20mph, participating drivers were first instructed in the use of the vehicle and their reactions measured in response to instruction to switch to driverless mode and take back control of the car. The data collected from these initial trials will be used to provide the groundwork for the next stage of testing and trials in the project.

Tony Meehan (Consultancy Director, Atkins) says that in the long term AVs can bring great benefits, but getting the user-interface right is key. He said, “Autonomous vehicles are going to be a global phenomenon and it’s absolutely critical that the UK and the West of England economy are able to benefit from this new global industry.

“Developing the technology and the skills needed to introduce something as radical as autonomous vehicles is crucial, but being at the forefront of this huge challenge is also about regulation, insurance, safety and maximising the benefits of autonomous vehicles. This project is trying to plug the gap between technology, the road network and the user.

“The research is looking at how users interact with an as yet unknown piece of kit – how they react when confronted with different situations and when confronted with a piece of kit that looks and feels very familiar but is really radically different.”

Professor Graham Parkhurst (UWE Bristol) says the social and behavioural aspects of how AV technology is introduced are crucial to making a success of the technology’s potential.

He said, “Reaping the benefits from this new technology depends on how we introduce it and how we decide to use it. We need to understand the human and social interface in order to fully benefit from this as a society.

“For example, a scenario in which all vehicles were replaced with AVs wouldn’t help traffic reduction and emissions reductions might be very small. An AV might drive more smoothly than a human driver but major benefits in terms of sustainable mobility will only be possible if we use these vehicles in a different way.

“So the research needs to help us understand if this technology could encourage people to travel in a different way. Is an AV more likely to be acceptable as a shared vehicle if none of the users is responsible for driving it? The promised benefits from AVs cannot be realised without understanding the social and human behaviours involved.

“Completely autonomous vehicles may be coming, but we are currently at the stage of ‘driver-assisted systems’ which are in use by some vehicles on our roads. In the short term drivers might benefit from better driver assistance technologies, but currently all vehicles on the UK network must have a driver attending to the situation and able to take control immediately.

“The really interesting questions around AV relate to transition. Many people might eventually like us all to end up with fully autonomous cars, but as the technology emerges politicians, transport planners and road engineers will need to deal with a future of mixed conditions where the road network will have to cater for AVs as well as other road users.”

Tests next year will look at different road conditions, the AV emerging from junctions, manoeuvring around roundabouts, and choosing appropriate gaps in traffic that make the occupants feel safe without unnecessarily slowing down traffic or adding to emissions.

Venturer is jointly funded by government and industry. The competition was delivered by the UK’s innovation agency, Innovate UK.

For a full list of consortium members and further details, visit the Venturer website.

Handover – A key issue to facilitate take-up of the technology in the real world

One objective of the trials currently being undertaken by the VENTURER consortium is to generate a better understanding of how drivers manage the handover process. Addressing the behavioural and physiological constraints around handover is vital to developing a proposition which meets user expectations and maximises the potential safety benefits of autonomous vehicles.

Handover (for these purposes referring principally to the transfer back to the driver of control following a period of autonomous operation) is one of the main issues to get right. Previous studies have demonstrated that the switch from autonomous mode is a key area, particularly in terms of the driver regaining lateral control of the vehicle (see the VENTURER literature review published earlier this year for more details).

Identifying the best way to manage the handover process will therefore be crucial to realising the potential of the technology. At a fundamental level, failing to achieve a safe handover process is likely to bring unacceptable safety implications. The technology cannot make driving less safe than it would be without it.

Users may also conclude that they are not comfortable with the driving experience if anticipating the next occasion on which they must resume control makes driving more difficult or stressful. The handover process has to offer a user experience that drivers want.

Drivers may also not embrace the technology if they have concerns about the liability implications of handover. Will a driver want to risk not being able to resume control at short notice and being (unfairly) held liable in the event of an incident?

OEMs, insurers and Government therefore all have an interest in understanding the parameters for handover and ensuring that an appropriate liability model is in place. As ever, risk and benefit will as ever be at the heart of the solution, with the legal and regulatory model becoming an integral part of the system.

Authored by:
Edward Barratt
Senior Associate, Burges Salmon LLP

How should local authorities prepare for the arrival of CAVs?

There is no question that cars are getting smarter but the real question is how smart can they get and when is this likely to happen? Connected & Autonomous Vehicles (CAVs) will transform how we use transport, how vehicles interact with each other and how they interact with ‘network operators’ or ‘infrastructure.

Vehicles and highways are getting smarter and the two are interacting with each other more and more. But an important point to note is the distinction between connected and autonomous vehicles. There are already a lot of vehicles operating on UK roads that are connected and some are even demonstrating semi-autonomous capability but it’s the seamless connection between the two, on the same road network, using the same infrastructure that poses the greatest opportunity and challenge for local councils.

Historically, roadside infrastructure has been fairly ‘dumb’. A junction would iterate through its cycle: green-amber-red-amber-green and so on, day after day. The addition of inductive loops and controllers at junctions allowed these locations to have some ‘awareness’ of the traffic approaching them and as a result they started to be able to modify their cycles in response to demand.

Urban Traffic Management and Control (UTMC) began in the late 1990s and is now the norm for most large conurbations. This means that a central control room has the ability to know the current state of all the main roads in an area and also to remotely update traffic timings to deal with the changing state of the roads.

Until now the main way in which a control room has interacted with vehicles is via drivers by, for example, sending messages to Variable Message Signs (VMS). This is changing, as more and more data about current traffic speeds is being harvested either through telematics units, mobile phones or Automatic Number Plate Recognition. This trend is likely to accelerate such that vehicle manufacturers, data aggregators and network operators have an increasingly detailed picture of the state of the highway network.

It is likely that drivers and vehicles will increasingly have much more accurate knowledge of the state of the road ahead of them, allowing them to reroute in real time but much more intelligently than a current in-vehicle sat-nav might (i.e. by sending vehicles over 7.5 tons one way and lighter/smaller vehicles in another). This poses the question: what will be the role of network operators such as local councils in the future be and how will we ‘manage’ traffic in a tangible way?

Two visions are possible: one ‘laissez faire’ scenario would see drivers and vehicle manufacturers leading the way in terms of choosing how to use data to maximise benefits to the occupiers of the vehicles; another more ‘interventionist’ approach could see network operators changing parameters (such as speed limits and priorities) in real time such that the road network as a whole runs more smoothly, reducing the feeling of driving between semi-urban (less congested) and urban (more congested) roads since the traffic volume ‘feels’ the same . The latter vision is more like an air traffic control scenario where the control centre knows about the location and movement of every vehicle in real time and can issue instructions about what it should do next.

How much control network operators will have in the future is a huge, complex and ethically-charged question. In order for these types of scenarios to become possible there will be a need for standard communication protocols, so that every vehicle can communicate unambiguously with the control centre, in a secure way.

Could vehicles in future be required to have a certain level of connectivity (and autonomy) before being allowed to enter certain areas? Who will control or prescribe how the vehicle operates within this ‘zone’? Could there be CAV-only lanes or CAV-only districts? Of course, the issue of handover is key to exploring this as well as the transition phase between standard driven vehicles and mainstream CAVs.

It is likely that different countries, different states and different cities will implement regimes that address their geographic needs, capability and capacity. Some – Singapore seems a possible candidate – may be more comfortable with fast implementation of compulsory standards. Others regions may take a more libertarian approach considering that intervening in a person’s right to travel where they want when they want is not acceptable and in these types of areas it is likely that CAV systems would be optional and so take up slower.

What is clear is that local authorities are not currently geared up to manage an ‘air traffic control’ level of intervention. If this is the way CAVs develop, then there will be a need for a large amount of coordination and investment in infrastructure to ensure that we are ready for the advent of mass CAV based transport. There could be huge benefits from these types of investments and they could have major impacts on how our urban landscape looks, feels and is managed.

Ultimately, the golden thread alludes to an increase in productivity (GVA), creation of new jobs and attracting inward investment, since people and businesses are able to move and work in a more fluid and easier way. We also hope through our involvement in Venturer, the Bristol/South Gloucestershire region becomes the UK leader in CAV R&D. Our soon to be published economic impact study aims to demonstrate how we might achieve this.

Co-authored by South Gloucestershire Council and Bristol City Council