Industry throttles down on self-driving car
At the Geneva Motor Show the electric car is everywhere. But when it comes to market-ripe examples of the much hyped self-driving car the place is utterly deserted. Industry seems to throttle down on the – private – autonomous car due to high costs and low public acceptance after over-mediatized accidents in the US.
In fact the majority of those mediatized accidents are booked to the account of Tesla with several casualties with drivers using the Autopilot function inappropriately. Tesla itself warns that the Autopilot is a driving aid of Level 2 autonomy, Level 3 at most in the near future, but not a fully self-driving tool yet.
Pushing to fast
On the other hand in 2017 four high level managers and a dozen engineers responsible for the Autopilot function quitted Tesla because they thought the company was pushing toward autonomous driving too fast. According to them, the program’s latest version at that time was released too soon to be 100% reliable.
Meanwhile Tesla made it impossible for the driver to take his hands of the wheel for longer periods of time, as the system will shut itself off after several warnings.
Only one deadly accident
The only deadly accident we know of with a quasi fully autonomous car, was the one in March 2018 in Tempe, Arizona. A Volvo XC90 from Uber with its autonomous driving software and hardware hit a pedestrian crossing the street in twilight. The 49-year-old woman, Elaine Herzberg, died of her injuries in the hospital.
Although the car detected the pedestrian, it didn’t brake as the emergency braking systems appeared to have been shut off, as the investigation proved later. The Uber employee behind the wheel didn’t react in time neither as she was ‘drowsing’.
On Tuesday the public prosecutor in Tempe announced that he sees no longer ‘a legal ground for judicial liability’ against Uber. Details were not given. The prosecutor wouldn’t exclude further investigations though into the role of the Uber driver. She could be prosecuted for ‘culpable homicide’.
Although just a week ago Daimler and BMW announced a historic partnership in developing their autonomous cars, there is little new to discover in this domain at the Geneva Motor Show.
The technologies that are currently under development will go into series production as Level 3 automation in 2021 in the BMW iNEXT, where it will also be Level 4 enabled for pilot projects.
Daimler AG has been working on development projects for production not only for specific Level 3 vehicles but also for Levels 4 and 5. 2019 will see the launch in San José, Silicon Valley, California, of its first pilot program, with Bosch, on self-driving vehicles (Levels 4/5) in urban environments.
Several hurdles to clear
Apparently car industry has become more prudent in predicting the advent of the autonomous car. “There are several hurdles to clear”, says automotive expert Thomas Morel of McKinsey to AFP. “Apart from the maturity of the technology, the capacity to mass-produce it, legislation to allow it, there is the acceptance of the public and the price one is willing to pay for it.”
Today you need more than 50.000 euro in autonomous driving hardware and technology to have the car drive itself. “Letting the driver take his hands of the wheel on a highway isn’t so far away”, Morel says. “But even in the most optimistic scenarios only 15% of autonomous cars will allow this in certain circumstances”.
To expensive for private use
End of February, Carlos Tavares, CEO of the French PSA Group and president of the European car manufacturers federation ACEA, admitted that the costs are too high for a autonomous car to be reasonable for a private customer. “It makes no sense to develop cars that are so expensive for private use’, he said.
Just to illustrated the complexity of the technology: Intel calculated earlier that an hour and a half of autonomous driving will generate 4 terabytes of data to be processed. That as much as 3.000 users on the internet generate in a whole day.
Automous shuttles and robot taxis
Tavares sees a market though for autonomous shuttles and robot taxis. “Vehicles of which costs of technology are shared” and in which PSA and all others are investing. He foresees for private cars, autonomous driving will be limited to more and more advanced driving aids. Fully autonomous will be reserved for public transport.
Xavier Mosquet of the Boston Consulting Group confirms that the majority of autonomous cars will be commercially used in the next 10 to 15 years. Private autonomous cars will remain marginal, due to their price.
Global management consultancy firm AT Kearney that questioned 150 industry leaders on the matter, foresees autonomous cars could have a beneficiary effect on the environment on the other hand.
Energy consumption could drop with 30% due to the fluidity of traffic and cars – communicating to each other (V2V) – anticipating to each other’s behaviour. And due to massive sharing. Owning a private car will lose its pertinence, AT Kearney believes.