Along with ride-share, the autonomous vehicle movement has been a key element of many smart city strategies. Google’s self-driving car shows up on the game-changer list of many an urban futurist around the world.
But there has been much debate about which party bares the liability in an autonomous vehicle accident. Is it the “driver” or is it the software, the hardware or the design? To address this debate, Volvo last year took the bold stance of taking “full liability” for crashes in its driverless cars.
The recent death of a Tesla owner driving in autopilot mode has raised even greater awareness of the pioneering nature of these autonomous vehicles. I dare not equate well-heeled Tesla owners with space explorers. But there are some similarities about how the same danger and risk are inherent in disruptive innovations…especially when Newton’s Second Law of Motion (force = mass x acceleration) is involved.
It is interesting to note that Google, Tesla, Volvo, BMW and the other intelligent vehicle manufacturers have taken slightly different routes to autonomy. All have one or more of the five levels of autonomy used to describe the evolution of these vehicles:
Google has opted for a longer-term evolution to a fully autonomous vehicle that focuses on the “driver off” features. Others like BMW are engineering new vehicles that straddle the “hands off” and “eyes off” stages. They all aspire to “driver-off” but their shorter term go-to-market will focus on some degree of driver involvement.
So how do smart cities balance the risks to drivers, pedestrians and property?
The answer becomes even more complex if we fast-forward 5 years to when ride-share and driverless vehicles are bound to converge. As proof-point, Uber started testing its first self-driving car in Pittsburgh, Pennsylvania, this May. There were also reports that Uber placed an order for 100,000 autonomous Mercedes-Benz S-Class cars earlier this year.
For urban transportation planners, there is some reassurance in the fact that 95% of traffic accidents are caused by human error. While it’s still too soon to extrapolate, one could argue that by taking the “human” out of it, there will likely be dramatic reductions of accidents using the technology embedded in driverless vehicles. Needless to say, non-human causes could increase dramatically.
Autonomous vehicle laws seem to be cutting it down the middle. For example in California and a number of other states, the “passenger” in an autonomous vehicle must hold a valid drivers license, and the vehicle must have a steering wheel and brake. This would mean that the Google self-driving car, which does not have a steering wheel or brake, would be restricted to certain permitted experimental areas.
Many smart cities see autonomous vehicles as mobility solutions for the handicapped, blind, elderly or others unable to get a driver’s license. So the usage restrictions reinforce that we’re still a long way from some of the most compelling uses of this technology.
Regardless, urban transportation planners will not be deterred by autonomous vehicle tragedies. As the installed base grows, studies will show that human-controlled vehicles will continue to have a higher accident rate than self-driving vehicles.
In the meantime, civic officials will “slow down for the car wrecks,” whether autonomous or human, but they will surely not apply the emergency brake on their beta deployments.