I knew that most cars were festooned with computers, but 75 ECUs is now the norm and we’re heading towards 100 – the Chevy Volt already has that number.
I’ve read elsewhere that a high-end luxury car might have as many as 10 billion transistors in its circuitry…
I was also staggered to hear that the computer code that runs the Volt extends to 10 million lines — as much as in the new Boeing 787 Dreamliner.
“Less of a car, more of a mobile data centre,” joked IBM engineer Paul Campion.
That’s too much code for a human team to programme, so GM and no doubt everyone else, has computers that generate the code for the ECUs that run today’s cars. Humans check the output and program the computers that program the computers, but otherwise it’s an automated process.
No wonder sharing electrical architecture is one of the main driving forces for increased platform sharing in the car industry.
With all these clever ECUs running the major functions in modern cars it really does seem a small step — by which I mean a decade or so — before some of the driver’s major decision-making is taken away by lumps of silicon.
Yesterday we reported GM’s plans to link these sensors together to make a ‘semi-autonomous’ safety system that will over-ride the steering and brakes in an emergency and I’m willing to believe GM’s prediction that it will become a production reality around 2017 – 2020.
I also didn’t know that there are already engineering standards in the US for vehicle-to-vehicle communications (V2V), which means the way is open for cars to communicate to each other by radio, so-called DSRC or dedicated short-range communication, leading to active crash avoidance systems.
And — I kid you not — experts are predicting a significant addition to the vehicle homologation process in which cars will be ‘licensed’ for their driving behaviour when the computers take over. That will solve the problem of who is to blame in the crash of the future — driver or car?
Incidentally V2V also opens up the possibility of junctions without traffic lights and swarms of cars zooming across each other’s path, avoiding each other rather like a pack of pedestrians on a busy pavement. Although I’m cynical about governments ever loosening their grip on the control thattraffic lights bring.
Of course, testing the feasibility of these technological breakthroughs is a global activity. In Japan, for example, Nissan has transmitted in-car anti-lock brake data to a central database to build a picture of slippery roads or tricky corners – information that can then be beamed to other cars as they approach the danger zone.
But what this conference was really trying to tell us is that technology is capable of letting cars travel safely, closer together at speed, to maximise the usage of road space.
When you consider the sobering prediction that the UK’s population will break through 70m in 2030 and match Germany’s somewhere in the late 2030s, this may well be the only way to keep Britain’s motorists mobile in the next 15 years — remember that’s just two model cycles in car-industry timescales.
Finally thanks to IBM speaker Paul Campion for this gem: the transistor was only invented in 1948, yet it’s reckoned that today, for every person in the world, there are an astonishing one billion transistors.