Autonomous vehicles and the future of driving
Every day, we hear about future tech that may or may not succeed. The main shift in the car manufacturing world is that companies are now being valued on how autonomous they are or have the potential to be.
The future is finally here. Nearly.
Separating the hype from the reality
Only now are we starting to get through the hype cycle around autonomy. What is this product really? How realistic is this? The answer is still unclear but we're beginning to understand the boundaries of autonomy better.
When a manufacturer says "All our vehicles will be 90% autonomous by 2021," what does that missing 10% represent?
It may mean that the car is autonomous for the motorway part of the journey, which is fairly safe. A brand-new motorway will have defined white lines and clear signage, making it easier for the car to keep on track. But over time, as weather leaves road surfaces uneven and lines begin to fade away, autonomy is going to have to work harder.
We face so many uncontrollable situations every day that will weaken autonomy, like bad weather interfering with cars sensors and requiring take-over. If a human needs to intervene, it isn't true autonomy; it's partial.
Another practical concern is processing power. Keeping hundreds of thousands of cars running with all their added autonomous features demands a lot of power and we don't yet have the infrastructure that would allow full-scale take-up, let alone with 'clean' energy.
The power has to be sourced from somewhere and it seems we're a long way off from an economical solution.
The young driver market is an area where autonomy hasn't been properly discussed yet. Most media reporting has gone straight to full autonomy, with the assumption that young drivers just won't have to drive in the near future.
But it's the inbetween that's concerning. Not all drivers will have an autonomous vehicle straight away, if at all - so we'll have at least a decade of mixed-autonomy vehicles on our roads, which causes issues in itself.
If a young driver in a 2008 Corsa hits a brand new autonomous vehicle with all the latest tech features, that's going to be a very expensive claim. A bumper that used to cost £100 to replace could cost more like £5,000 as it's now filled with sensors and radars. You won't just be able to pop down the garage and be back on the road within a week.
Then there's the mix of autonomous and semi-autonomous vehicles. Some mainstream manufacturers believe the risk is higher with a semi-autonomous car; switching between relying on your own reaction times and relying on the car to react for you is confusing - especially if the driver is tired or regularly drives different vehicles.
If a new driver starts off in a semi-autonomous car with an autonomous braking feature, for example, that's great. It's safe and means that the car is reacting to a hazard much quicker than a human ever could. But then what happens if that driver has to jump in a different car as part of their job, or has a crash and is given a replacement that doesn't have this feature? Complacency brought about by the new feature would have already set in and it could end very badly.
Safety comes first
Vehicle autonomy exists predominantly to make things safer. Whether that's collision avoidance systems or lane detection sensors, the end goal is indisputable: reducing crashes on our roads.
A machine is quicker than a brain because the road is all it's thinking about. We all believe we're concentrating on the road but we're not. Not by the standards of AI.
Humans are trying to remember to get a birthday card for Mum and figuring out what to have for dinner. An autonomous car just does the driving.
Autonomy has huge potential to eliminate the dangers of driving and now implementation is actually happening, it's definitely the right time to get interested.
Huge improvements still have to be made before autonomy is rolled out to the masses. Vehicle-to-vehicle and vehicle-to-infrastructure communication, mapping, hacking...
Full autonomy is the dream and great leaps are happening to get there, but we're still some way from it being reality.
By Selim Cavanagh
Selim joined ingenie in 2018 as CEO. His first car was a brown Rover 216 GLi and you can follow him on LinkedIn.