Self-driving cars and the role of the driver
A study carried out at Stanford University concludes that transitioning from autonomous driving mode to person-driven mode when a vehicle is in motion poses numerous adaptation problems for the driver and should not simply be preceded by a warning to take responsibility for the vehicle. The so-called handover has been questioned for its safety and for creating a dangerous window during which the chance of an accident occurring increases significantly.
The study questions the route taken by most companies in the automotive industry, including Tesla: the idea of introducing autonomous driving technology as an aid or assistance to driving only under certain circumstances, such as on motorways, in certain built-up areas or based on other parameters — including whether the driver is paying attention — that could affect the performance of autonomous driving systems.
On the other hand, the strategy pursued by Google, whose self-driving project could soon be ready to spin out from Alphabet could give it a huge competitive advantage: autonomous driving not as assistance, but as the total removal of all human intervention. The idea is that the only way to ensure transport safety is to remove from the equation we clumsy and unpredictable humans, responsible for the vast majority of accidents, and who are best relegated to the role of passengers. In short, by not having to pay attention to what is going on around us we can do other things during our journey time.
For the moment, assisted driving has the debate, based on the fact that, on the one hand, companies with a long tradition are already producing such vehicle, and on the other, that this incremental approach gives drivers time to adapt. But studies that highlight the dangers of handover could pose problems to the introduction of these systems: in Germany, Tesla cannot call its system Autopilot, even though the system requires a decreasingly lower level of human intervention.
Unsurprisingly, little is know about Apple’s plans, but so far there is no indication it has lost interest developing self-driving technology.
Were Google’s approach to be adopted, a company that has never considered manufacturing vehicles, instead preferring to license its technology, would potentially find itself able to sell its technology to other companies, which might justify a certain level of optimism in the transition from the experimental to the commercial phase of self-driving vehicles.
Any perceived failure of handover-based systems could delay many carmakers’ plans, possibly resulting in delays and significantly slower adoption rates.
Calculating possible scenarios in this complex area is difficult. That said, it seems clearer with each day that we are no longer talking about if, but when. It might be four, five or 10 years, but the self-driving car, under whichever variant or combination, is here to stay.
(En español, aquí)
