The UK government has said it believes that driverless cars and other light vehicles will be on Britain’s roads by 2021 – two years from now. Elon Musk has also recently made the bold claim that within a year there will be a million Tesla vehicles on the roads, mainly in North America, will fully operational and perfectly safe self-driving hardware and software. That would mean the only thing holding the company back from flicking the switch that would update and activate that software would be regulations catching up.
However, a recent survey conducted by Forrester, the market research and analysis company, shows that industry experts are more conservative. The survey, which polled the opinions of 54 automotive industry experts, showed that they believe it will be at least another 10 years, possibly more, before we are truly ready to embrace a driverless future.
The experts believe the hardware and software that will enable safe driverless vehicles is still to be really perfected, cyber security of driverless vehicles is still a serious threat and that they will be too expensive for the mass market until economies of scale are sufficient to bring prices down to around current price points. The Forrester research was commissioned by Arm, the British-based microchip manufacturer now owned by Japan’s Softbank.
Robert Day, Arm’s director of automotive solutions and platforms, responded to the research with the conclusion:
“My conclusion from reading the Forrester report, and what I know, is that fully autonomous vehicle fleets will happen. The deployment of the first truly self-driving cars is likely to be restricted, such as cars operating only in pre-defined and well-mapped areas, in certain environmental conditions, such as daytime and in dry conditions, and at limited speeds.”
Before that can happen the report suggests there are several major obstacles that still need to be overcome. Firstly, there is still serious scientific debate around which components of driverless control systems are non-negotiable technology if driverless vehicles are to be safe.
The most contentious question in this area is if Lidar light sensor technology is a necessity. Lidar systems send out pulses of light which build a detailed 3D map of a car’s surroundings. Tesla boss Elon Musk has referred to the technology as a ‘fool’s errand’. He believes the technology is both too expensive and bulky to be practical and Tesla cars shun it in favour of a combination of cameras, radars and ultrasonic sensors. Waymo and Uber, two of the other companies at the vanguard of driverless technology and systems are both building their driverless OS around Lidar.
The cyber security of driverless vehicles will also have to be proven as airtight. Autonomous vehicles will be connected, which means it is paramount that software does not contain any backdoors that could conceivably allow hackers to gain control. Security problems could, however, be much simpler than that. Chinese security testers last month claimed they were able to confuse Tesla’s software into turning the vehicle into oncoming traffic with nothing more than brightly coloured stickers placed on the road. Tesla’s response was to dismiss the test scenario as ‘unrealistic’ but would that rebuttal be enough to convince potential driverless passengers?
If ill-intentioned groups or individuals are prepared to be suicide bombers or drive vehicles into crowds, would they not happily place brightly coloured stickers on roads frequented by driverless vehicles? It’s obvious that driverless vehicles cannot be so easily confused if they are to be given permission to take to the roads.
The Forrester report also cast doubt around the manufacturing scalability of the currently expensive and energy hungry driverless systems of computers and sensors.
Arm and Mr Day’s final conclusion was that it will take another decade at least before driverless fleets are ready to be mass produced and let out of controlled zones.