A Sober Second Look At Self-Driving Cars
While TTAC‘s Mike Smitka published an essay urging readers to reign in their expectations regarding autonomous cars, a new report by MIT’s Technology Review pours even more cold water on the utopian fantasies of those waiting for the day when humans are no longer in control of the automobile.
While the full text is available at MIT, the American Enterprise Institute summarized the obstacles faced by autonomous cars in a series of handy bullet points
- The self-driving car can’t drive itself in 99% of the country.
- It knows almost nothing about parking, and can’t be taken out in snow or heavy rain.
- If a new stoplight appeared overnight, the car wouldn’t know to obey it.
- Google’s cars can detect and respond to stop signs that aren’t on its map, but at an unmapped intersection stop sign the car wouldn’t know what to do after it had stopped, and would probably remain stationary until a human driver intervened.
- The car hasn’t yet tackled big, open parking lots or multilevel garages.
- The car’s video cameras detect the color of a traffic light, and they’re still working to prevent them from being blinded when the sun is directly behind a light.
- Pedestrians are detected just as moving, column-shaped blurs of pixels—meaning that the car wouldn’t be able to spot a police officer at the side of the road frantically waving for traffic to stop.
- The car’s sensors can’t tell if a road obstacle is a rock or a crumpled piece of paper, so the car will try to drive around either. The car also can’t detect potholes or spot an uncovered manhole if it isn’t coned off.
Given all of the breathless hype regarding the technology, and Google’s introduction of their own prototype, sans pedals and steering wheel, it helps to have a contrarian viewpoint to dampen some of the exuberant enthusiasm professed by many who are better versed in the tech side of things, without understanding the unique subtleties of the auto world.
More by Derek Kreindler
Comments
Join the conversation
These are hardly show-stoppers. They are merely a few more items to include in the final solution. The future is already here, it's just not evenly distributed. It's silly to think autonomous cars will suddenly appear everywhere at once. They will appear on the Google campus & other controlled locations, then expand. We already have autonomous cars--self parking, radar cruise, lane-keep, auto braking, etc. these features will simply continue to get more inclusive, and human involvement will continue to decrease. As cars get smarter, we will build smarter infrastructure: V-to-V, V-to-I. Car can't 'see' a new stop sign? Well that new stop sign will be smart so that the car will automatically work with it--same with lights and parking lots/garages. It will be a transition, but it will happen.
Pretty much all of that is fixable with software and infrastructure improvements. Bitching that it can't park? I mean come on. We aren't even looking at the model t here, we're looking at a prototype of the prototype. It's early days yet, but I'm quite sure that a computer can and will be a much safer driver than the average driver. The computer never gets distracted, drunk, sleepy, or old. Automated cars would save tens of thousands of lives every year in the US. This is like saying personal computers will never be big because the Atari 7200 just isn't that much better than a typewriter.