Semi-driverless cars will seen in Europe early next year.
On a gloriously sunny day last week in Yokohama, near Tokyo, Nissan prepared to show off its latest driverless technology in an electric Leaf car. But only a few minutes into the demonstration, the vehicle came to a juddering halt.
“We’ve realised once again the difficulty of an unmanned system,” said Nissan’s vice-president Kazuhiro Doi apologetically, as engineers admitted the device may have been affected by the mobile phone signals of the watching media.
The incident demonstrated again that driverless systems remain far from perfect, a particularly important lesson for Nissan as it prepares to roll out partially autonomous systems on European roads for the first time in the new year.
The company is already selling “Pro-Pilot” systems — which allow the car to control acceleration, braking and steering in single lane traffic — in the Serena minivan in Japan.
But a much bigger audience — and test — awaits when it starts to include the technology in the next version of the Qashqai SUV for the European and international market next year.
It is the first time this level of automation, which is only featured in upmarket brands such as Mercedes and Tesla, will be available in mass market vehicles.
So far in Japan, 70 per cent of Nissan Serena buyers have opted to pay more than $1,000 extra for the added autonomous features. But examples of accidents involving other car brands with comparable technology — notably Tesla — mean there are still questions over the safety of a system that can drive itself for part of the time.
There are “valid concerns”
Such concerns will inevitably limit the potential market for Nissan — and rival groups pursuing similar partial self-driving technology such as General Motors. Indeed, GM has already pushed back the launch date of its comparable Super Cruise system until it is sure that it is fully safe.
Mike Ramsey, an analyst at Gartner, says there are “valid concerns” about semi-driverless technology, and that the likelihood of accidents involving autonomous cars will rise once more are on the road.
“My view is there may be ‘cracked eggs’ — accidents that should never have happened,” he says. “The chances of some accidents happening are higher if you have mass proliferation.”
The approach has also been deemed fundamentally unsafe by some rivals in the industry, such as Ford and BMW, which are pursuing full autonomy without intermittent steps.
But some car experts argue that semi-autonomous cars are an important innovation in their own right, helping cut out the need to drive in traffic jams, for example, and potentially increasing safety given the car, unlike humans, will not get distracted by children or mobile phones.
Nissan, for its part, insists it can make the systems safe. “Our autonomous driving is based on the safety technology that we have over the past years such as braking control [and] of lane departure alerts,” says Takashi Shirakawa, Nissan’s research and development director.
“It’s not a completely new technology, the foundation is already established and has been proved.”
The fully autonomous technology demonstrated by Nissan in Japan is intended for use in warehouses for towing finished cars on trailers, and is a long way from being placed in the hands of consumers on the roads.
But the building blocks are the same. The towing system installed in the autonomous Leaf car uses cameras and laser scanners to detect obstacles, kerbs and lane markings that are matched with map data. The new Qashqai will also contain scanners and sensors to monitor traffic, read road signs and control braking and steering.
In the first stage, it will only drive in a single lane on motorways. The next iteration, earmarked for 2018, will see the cars able to change lanes and overtake, while a more ambitious step will see the vehicles able to negotiate busy city roads with pedestrians and intersections by the end of the decade.
In all, Nissan plans to launch 10 vehicles fitted with automated systems by 2020.
“We will not launch this technology until it is ready,” says Paul Willcox, chairman of Nissan Europe. “The technology has to be fundamentally robust and reliable.”
Nissan also insists that drivers will continue to be in charge. Even when it is self-steering, drivers are told to keep their hands on the wheel at all time. The system beeps if you take your hands off, and eventually slows to a stop if no human response is detected.
The package also comes laden with warnings of the limits of the technology, calling itself an “assistance” service rather than an autopilot system.
But even Nissan admits that the technology is not flawless. Crucially, the car cannot rely on handing back control to the driver in an emergency. Even sitting behind the wheel, the human’s brain may not be sufficiently engaged to avert a crash.
“You can have your hands on the wheel, and feet hovering over the pedals, but if your conscious[ness] is beyond the operation, this is quite a dangerous situation,” admits Mr Shirakawa, adding that this interaction is more important to future iterations of the technology than driving in a single lane.
“We have not thought this out yet,” he adds.
Gartner’s Mr Ramsey says it “is incredibly difficult to answer” how the emergency handover will work in practice.
“The cars must be able to handle the emergency situation and expect that the driver will not come back,” he says.
But even with such questions outstanding, he expects companies offering such features will have a head start over those who have pursued more cautious approaches.
These systems are also, even in their current state, safer than human drivers, he adds. “If we actually followed the rules we would be extremely safe. The problem is we don’t follow the rules.”