‘Hands off’ – research into self-driving cars accelerates

You step into the driving seat of your car then shut the door. Turning on the engine, you gather your notes for the meeting ahead, a 45-minute drive away. You make a call back to the office, asking for the minutes from last month’s management conference to be emailed over. You sit back and relax, whilst scanning through the notes. “You have arrived at your destination” announces the on-board computer. You are using a self-driving car. This could be you, much sooner than you think.

Driverless technology gathers speed
Advancements in driver assistance systems has gathered pace over the last 40 years. Most cars now come with automated features as standard, including anti-lock brake systems, cruise control and collision avoidance. Now, the focus is on self-driving technology. Google has already built prototype vehicles and Tesla and Uber are also experimenting with automation. Self-driving car concepts dominated this year’s Geneva Motor Show with models showing what the future of self-driving could look like. There is a lot of excitement around this technology, and rightly so – for it may have huge potential to help those people who can’t drive: the elderly, disabled and those without a driving licence. However, the most eagerly anticipated benefit is that they reduce the number of vehicle accidents by eliminating human error.

Safety and driver assistance systems
A timeline of how driver assistance systems have evolved:

  • 1971 – ABS prevents wheel lock-up to maintain friction between wheel and road surface.
  • 1995 – Electronic Stability Control maintains directional controllability through brake intervention.
  • 1999 – Adaptive Cruise Control (ACC) replaces the driver in the car following task by driving at a set headway.
  • 2000 – Lane Departure Warning
  • 2005 – Blind Spot Warning to assist in lane changing.
  • 2006 – Lane Keeping System applies corrective steering responses to steer the vehicle back into lane.
  • 2006 – Forward Collision Warning
  • 2007 – Collision Warning with Auto Brake intervenes in potential rear-end collisions if the driver does not react.
  • 2007 – Stop and Go extends ACC to stop and start traffic.
  • 2010 – Volvo Cars Pedestrian Detection with Full Auto Brake.
  • 2011 – Mercedes Active Lane Keeping Assist applies brakes if it senses the vehicle is drifting out of lane.
  • 2013 – Volvo Cars Cyclist Detection with Full Auto Brake.

University of Leeds in the driving seat
Leeds is a leader in this field of research. Our Driving Simulator – one of the most advanced driving simulators in a research environment – is a great tool for understanding how people behave whilst in a self-driving car on a ‘real’ motorway. Academics at the Institute for Transport Studies (ITS) are researching safety, human factors and the effects this technology may have on our environment.

Safety first
Oliver Carsten, Professor of Transport Safety thinks that self-driving vehicles would save a lot of crashes by cars on the road ‘communicating’ with each other and reacting faster than a human could in the event of an impending incident – “We wouldn’t have problems with distraction, fatigue, speeders, drunk drivers or drivers under the influence of drugs.” However, Tyron Louw, a human factors expert working at the University on understanding driver interaction with automated vehicles, thinks that full automation is a long way off. “For the foreseeable future, automation technology in vehicles will still require the driver to take control in some situations, and because we know driver error has some role in the majority of crashes, we need to look closely at that interaction,” says Tyron. “It is our goal, as human factors experts, to understand drivers’ capabilities and limitations in this new environment, because automation is marching towards us and humans aren’t going anywhere.”

Human factors
Extensive research is being conducted by Leeds academics into human factors, looking at whether humans are able to intervene if something doesn’t go to plan in a fully autonomous self-driving car. In a Leeds TEDx presentation, Tyron Louw explained why research in this area is particularly crucial to the technology’s implementation: “Little attention is given to how it will affect us as drivers and whether we are actually ready for it,” said Tyron. “It’s not only about whether we are happy to use a self-driving car, but whether we are capable of using one?” As automated systems become more and more sophisticated, potential problems with how we interact with them will naturally evolve. However, if there is an understanding of how and why humans interact in the way they do, then these problems may be diminished and car manufacturers can be more informed when designing these systems.

It is easy to assume that the only group of people to study when researching and even talking about self-driving cars are the people in the vehicle, but cyclists and pedestrians also need to be considered.  Dr Natasha Merat, Associate Professor, has been conducting research in this area and has collected data on what pedestrians and cyclists want to see happen with self-driving cars. As part of a CityMobil2 project, Dr Merat has conducted surveys to understand if these groups can share the road with self-driving cars – particularly in relation to low-speed vehicles. Whether pedestrians and cyclists will be able to operate on roads where self-driving cars run to the speed limits in the future needs more research.

Dr Merat heads the Institute’s Safety and Technology research group which focuses on new technology and how road users can benefit from that technology. The group is steering the focus onto human interaction, performance and behaviour when using vehicles at different levels of automation.

The University’s Driving Simulator helps improve understanding of human interaction. Researchers can manipulate a scenario in the simulator and compare human reaction when driving normally and when in a self-driving car. Current findings show that in the majority of situations, human reaction time is slower when someone is in a self-driving vehicle and suddenly has to switch it to manual drive, than if they had to react to a situation when driving normally. For Dr Merat, this could be due to a number of reasons: “Firstly, they may assume the automation has driving matters under control and as a result become complacent. Or, they become distracted whilst concentrating on other tasks and by the time they realise they need to resume control, it is too late.”

Watch the human response when a self-driving vehicle encounters a hazard:

[Video: this experiment demonstrates that extensive research into human factors is needed and is a prime example of the risk of people become too trusting of computerised driving systems.]

Helping the environment?
There’s no doubting that, without human intervention, self-driving cars could optimise energy efficiency, but will they really bring environmental benefits? Alongside US academics, Dr Zia Wadud, who holds a joint appointment with ITS and the University’s Centre for Integrated Energy Research has recently looked into such effects. Dr Wadud found that the introduction of self-driving cars would increase car use, resulting in an overall 2-10% increase in energy consumption. Increased car use would come, in part, from inviting those who cannot drive a manual car onto the roads, such as the elderly and those with disabilities.

However, Dr Wadud does recognise some potential energy benefits. If there was widespread adoption of self-driving cars on the road, cars could communicate with each other to create a “platoon”, meaning that vehicles would face less air resistance and therefore deliver a 4-25% reduction in energy consumption. And if vehicles can interact with each other and road infrastructure, such as transport control systems, this will lead to a smoother traffic flow, meaning less congestion on roads and an energy use reduction of up to 4%.

Planes, trains or self-driving automobiles?
So what would happen to public transport that so many of us currently use? Dr Wadud has looked into what other factors people find important when travelling, for example, time efficiency, and applied this in the context of self-driving. “When you make a decision about transport, you also take into account non-financial costs.” Dr Zia Wadud. One of the current advantages of public transport is that it not only allows you to relax, but also to use your time doing other things, such as reading and work. But what if you could do these things in a self-driving car, whilst also enjoying privacy and convenience? “That changes the whole equation,” says Dr Wadud. His research shows that these attractive benefits of a self-driving car could result in an increase of 60% in car travel in the US. Although it might be a more enjoyable ride for us, this could have damaging consequences for those with jobs in the transport industry. Once self-driving cars are on the road, it is only a matter of time before other modes of transport – including public transport – will also become automated. Indeed, there are already automated buses in Europe.

Sharing culture
Another area for consideration is whether people will become more concerned about the technology and software in the self-driving car than the hardware. If people are less interested in the latter, this could reshape the whole industry model and prove a threat to manufacturers. People may also begin to feel less emotionally attached to cars that they are not driving themselves, making less sense to have one of your own, leading to a car-sharing culture. Professor Oliver Carsten identifies that this poses a huge challenge for manufacturers, because it is likely that people will end up not wanting to own cars but just want to use them. Dr Zia Wadud believes sharing offer a large potential to reduce the environmental and energy impacts. He suggests that in a sharing or automated taxi model, the size of the self-driving car can be matched with the trip and occupancy type: a small car used for a one-person commute, and a larger car for a family trip. If this was taken on, it could reduce energy demand by 21% to 45%.

The road ahead
Before self-driving cars enter our roads, there are still many questions that need to be answered. How will our road regulations and laws change when self-driving cars are introduced? Do people even want to use self-driving cars? Questions surrounding driver liability have also been sparked following a recent collision involving a Google self-driving car and a bus; the vehicle assumed the bus would slow down to let the car pull out – which it didn’t, resulting in a crash. Who is to blame? “If it’s a truly self-driving, fully automated car,” says Professor Carsten, “then the car is responsible.” Volvo Cars has announced that it will accept full responsibility for the actions of its self-driving cars when in Autopilot mode – potentially starting a trend for other manufacturers to follow. As self-driving technology advances, legal experts will have to look more closely at important issues of liability and responsibility if the technology is to truly become a part of our transport system. So when can we expect self-driving cars to be on our roads? Dr Natasha Merat thinks it depends on what level of automation these self-driving cars possess and believes that we are 20 years away from full automation, whilst Professor Oliver Carsten believes we could have some self-driving cars on our roads in less than 10 years.
In other words, don’t take your hands off the wheel just yet.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s