Autonomous vehicles move ever closer to becoming reality, and car-makers have slowly been introducing more sophisticated systems in the last few years. As cars take more and more control of the actual driving and remove control from the driver, so the question of liability in the case of an accident arises.
Tesla recently released a software update for their Autopilot system which enabled the car to drive autonomously in certain situations, but Youtube videos posted by Tesla drivers of the system in action have already highlighted cause for concern. Tesla expected its customers to apply common sense and stick to the general rule of keeping their hands on the wheel, but videos have shown drivers not only removing their hands from the wheel, but also using the system in inappropriate places.
This has brought sharply into focus the question of what happens when the car is responsible for an accident – which has long been seen as one of the big hurdles of bringing autonomous tech to reality.
The Question of Responsibility
Current estimates suggest that over 90% of all fatal road traffic accidents are caused by human error, and while Google for example say that their technology is almost on a par with a human driver, it still leaves a rather large 10% of grey area. What happens if the car is responsible for an accident? What happens if the software is hacked and subsequently fails? These are questions that OEM’s and legislators must get to grips with before fully autonomous cars will ever see the light of day.
In recent weeks Volvo, Mercedes-Benz and Google have all said that they would bear responsibility if an autonomous car caused an accident due to the technology. That’s not to say that those companies expect accidents – far from it – but it is a logical step towards reassuring the public and government regulators of their confidence in autonomous technology.
Tesla Plans for Autopilot System
Persuading members of the car-buying public that self-driving cars are safe will be one of the big hurdles for the auto industry to overcome, and advertising campaigns are already well underway to ease the approach. However, as Tesla recently discovered, people are inherently unpredictable and can’t necessarily be relied upon to utilize the technology in the best or most applicable ways.
Since releasing th software update that enabled the model S to drive autonomously in situations such as on a multi-lane highway, several owners have released Youtube videos of Autopilot, some of which have shown drivers using the tech in the wrong kind of situations, or removing their hands from the wheel inappropriately.
This has the potential to be a PR disaster for Tesla, and CEO Elon Musk summed up the company’s position by saying, "This is not good. We’ll put on some constraints on Autopilot to stop people doing crazy things with it."
In reality, it was entirely predictable that some people would misuse the technology, and Tesla may consider a similar system to the Mercedes-Benz Steering Assist, which features sensors to detect if the driver’s hands are on the wheel. What the situation also highlights is that people do not always follow predictable, logical paths and the autonomous car must be intelligent enough to accurately predict human behaviour when it encounters other drivers, cyclists and pedestrians.
Dealing with Pedestrians
Accounting for this unpredictable behaviour is something that Nissan have been working on, and they unveiled the new IDS Concept at the recent Tokyo Motor Show.
The concept embodies Nissan’s vision of the future of autonomous driving and zero emission EV’s, and they released a detailed promotional video alongside to highlight the sophisticated vehicle management system. Understanding how vehicles can communicate with other people as well as the driver is central to the concept, which has a tagline of ‘Together, we ride’.
Bu way of example, there are situations when it is polite to let a pedestrian cross a road, but other situations where it may be preferential to continue to prevent halting the flow of traffic. The autonomous car of the future, says Nissan, should be capable of assessing what to do in these complex situations, and also be capable of communicating its intentions to passing pedestrians, cyclists and other road users.
To achieve this Nissan have used a series of exterior lights and displays to convey the car’s awareness of its surroundings to pedestrians and other road users, and to signal its intentions. The LED strip on the side of the car’s body shines white when a pedestrian or cyclist is nearby to show that it is aware of their presence, while another display facing outwards behind the instrument panel can flash up messages such as "After you" to signal to a pedestrian that it is safe to cross the road.
Nissan plan to introduce the technology in a number of stages, but say that fully autonomous tech could be ready by 2020, with the caveat that it will be subject to legislators allowing such features.
Autonomous Tech Coming to Market
The race to produce the first autonomous vehicles is increasing in pace, with Volvo and Toyota both recently releasing details of their latest research and developments in the area. Meanwhile Mercedes-Benz has shown off a production-ready truck fitted with Highway Pilot, which enables it to drive autonomously on highways. The truck drove itself on a test drive on the German Autobahn from Denkendorf to Stuttgart airport in early October; although at present regulations only allow the truck to drive autonomously in ‘test’ phase.
While the development of autonomous driving has always been about advancing the technology, attention is now turning towards the legislative and logistical problems which will need to be ironed out if the ambitious promise of self-driving cars is to become a reality by 2020.