Self-Driving Cars: Inside the cabin

The above talk was given at SXSW this year. It’s an excellent talk, but I understand if you don’t want to watch the whole thing.

I especially love how he totally doesn’t mention that the “driving on the freeway” concept is exactly how Tesla’s AutoDrive currently works.

The bit that really caught my attention was at 17:09 where Chris Urmson shows a rendering of what a self-driving car’s interior could look like. No steering wheel, a small display… more like a small living room if the chairs or chesterfield had seatbelts.

A rendering of what the interior of a self-driving car might look like. Instead of a steering column and control surfaces there is a luggage shelf and a small display.

Something in my brain was instantly repulsed by this. Not the design, which is fine. Not the lack of leg room, though my lower back tightened up slightly. Not the complete lack of heating vents, which would make it useless in Canada.

Eventually I realized it was, incongruously, the absence of a steering wheel that caused me to go “Nope.”

But this is a self-driving car! The whole point is that there is no steering wheel! Don’t you get that?

Well, yes, I do. Which is why it took me so long to figure out what was bothering me about the render. I was, and still am, convinced that the lack of a steering wheel and other control surfaces is a benefit, not a detriment. But there are some use-cases I think a lack of a steering wheel will significantly hamper.

Self-driving cars are excellent if the car knows where you are and you can tell the car where you are going. “OK Google Car, take me to work” “OK Google Car, let’s *sigh* go to the in-laws”

But what if you don’t? “OK Google Car, take us someplace nice for dinner”

It can take you to the closest Google+ listing for a restaurant. It can find the most efficient route to take you to a Michelin Star-rated eatery in a neighbouring metropolis. But you can’t browse. You can’t see the line from the street and change your mind and say “Actually, where else could we go?”

But maybe this isn’t a common enough use-case to care about. Maybe having to choose from amongst the available options before you put the car in motion is a good thing.

“OK Google Car, take me to my coworker’s BBQ” << Where is that? >> “According to the invitation, head to the Red Barn past the crossroads, take a left, then keep going until you see the balloons or a sign saying Kalamazoo”

Or: << Arriving at destination >> “OK Google Car, make sure not to park next to the begonias or my mother’ll kill me”

Or: << Arriving at destination >> “Aw nuts the parking lot’s full. Guess we have to park in the field. Watch out for the furrows or we’ll never get out.”

Is it enough to create a car that can only do most of what other cars can do? There is already an understanding of how that works in the snowier parts of the world: there are some cars that can drive before the snowplow gets to your street, and there are other cars that cannot. But will this sort of restriction, like range anxiety for electric cars, slow adoption of this crucial piece of transportation infrastructure?

I think it is necessary that people who cannot drive still be able to get where they need to go. I think it is necessary to eliminate traffic fatalities as an understood fact of life.

I think it is necessary that the Google Self-Driving Car team think some more about how the car interacts with its occupants at the same time they’re thinking about how the car interacts with its adjacent road users.