Never mind the ‘how’ of autonomous driving – let’s focus on the ‘why’
Picture credit: Volvo
If, like me, you remember the original release of Back to the Future II in 1989, 2015 holds a particular place in the automotive imagination. Come this year, it promised us, there would be flying cars, roads in the sky and the small matter of on-board nuclear fusion generators.
But while some of the predictions encountered by Marty McFly et al have come to pass (wearable technology, video calling, 3D cinemas), and others been outstripped by reality (the future McFly is famously fired by fax), the real world has rather lagged behind the film’s four-wheeled visions.
Yet there is one significant exception to this, where what sounds like science fiction is about to become reality: cars that drive themselves.
In 2015 we are at the point where the software, hardware and engineering expertise can equip a vehicle with the systems it needs not only to steer itself on a track but also to interact with a real-world environment, other road users and pedestrians.
And momentum is building. A UK trial in Greenwich, Milton Keynes, Bristol and Coventry, backed by the British government, was confirmed earlier this year. Google announced last May that its prototype is now fully functional. And already in Sweden, we at Volvo already have self-driving cars on the road in Gothenburg ahead of a trial that will put 100 cars in public hands for a real-world test from 2017.
This is the beginning of the biggest revolution in transport since the Model T opened up car ownership to everyone. But as we enter this brave new world, there is one question we must keep at the forefront of our minds: why? Why do this?
Manufacturers and researchers need to be clear on the real-world benefits and be able to communicate them to motorists, government and the related industries that need to be brought on side, such as insurance. Without this single-minded focus, self-driving cars will remain a novelty and their development scattergun. We can see this already with the very different visions of what an autonomous driving car should look like and the technologies involved.
Self-driving cars therefore need to be related back to a real driver need and, for us, the biggest need is always this: keeping passengers safe. To that end, autonomous driving is not a side project at Volvo but an integral part of Vision 2020, our aim that, within five years, no one should be killed or seriously injured in a new Volvo.
Some may see a slight contradiction in the argument that letting cars loose on the roads without drivers will make our highways safer. But there is a clear logic to it: the biggest cause of accidents, by far, is driver error. So to make the biggest reduction in accidents we should remove driver error. And the simplest way to remove driver error is by doing away with drivers altogether.
Of course, this is not to say there should be no human input. Our expectation is that cars will be fitted with active and autonomous driving modes and require someone in the driving seat who is able to take the wheel if needed, which means they will need to be a qualified driver and in a fit state to drive.
A car can be much more alert, react more quickly and make better decisions. A case in point: the City Safety system fitted as standard to all Volvos keeps an eye on traffic, cyclists and pedestrians, and alerts the driver if a crash is imminent, or automatically applies the brakes if unavoidable. Through this it can cut accident rates by a quarter.
And if we’re going to view autonomous cars as a way to keep drivers safe, it follows that in order to deliver the biggest benefit we need to pursue the broadest uptake. That’s why Volvo’s Drive Me programme is being implemented with real, production cars, rather than the small, pod-like cars that are involved in some other trials. We believe self-driving cars should fit with the current model of car ownership, rather than requiring further investment in a second vehicle.
To demonstrate that self-driving cars can be part and parcel of daily life, we will also be running a real-world trial in Gothenburg in two years, working with the local authorities to put 100 cars driven by members of the public on real roads, mixing with other traffic.
This focus has led us to address the unglamorous practicalities head on. If we’re to make self-driving cars a reality we need not only to look at high-tech sensors and location services but create systems that are incredibly tolerant to faults and errors. It is relatively easy to build and demonstrate a self-driving concept vehicle – the tech-led approach – but to make a system that will work in the real world – the driver-led approach – requires a system that’s safe, robust, easy to use and affordable for ordinary customers. To that end, motor manufacturers need to take a similar approach to the aircraft industry: with multiple backup systems that ensure the car is able to drive itself safely if any part of the system is disabled for any reason. Being 99 per cent reliable is not good enough; we need to be much closer to 100 per cent.
Finally, having a clear vision for why we want to introduce autonomous driving helps bridge the gap between where we are and where we want to be. By making it relevant to a real driver need it makes it more likely that self-driving technologies can be adopted in current cars, providing a return on your R&D investment, paving the way step-by-step to full autonomy, and, crucially, improving driver safety.
A surprising amount of autonomous technology is already in use – modern cars have more of a mind of their own than many drivers might think. As well as the aforementioned City Safety, the new edition of the XC90 will park itself, the driver only needing to control the accelerator and brake. It is also the first car in the world to feature automatic braking if the driver turns in front of an oncoming car.
We are still several years from fully autonomous cars being available commercially. But by looking back to some core principals we pave the way to the future.