Flutterby™! : Self-driving cars in Nevada? 2011-05-12 02:03:59.994679+00

Self-driving cars in Nevada?

2011-05-12 02:03:59.994679+00 by Dan Lyke 14 comments

Google lobbying Nevada to allow self-driving cars:

Since the car drives itself without human input, Google is asking for an exemption that would allow occupants to text while sitting behind the wheel.

This transportation stuff I've been working on is interesting because there's going to be a big shift in how we think about automobiles shortly. To now, we view texting as a distraction from driving. In 20 years, we'll view driving as a distraction from texting.

#Comment Re: made: 2011-05-13 20:08:30.81155+00 by: jeff [edit history]

I'll keep my terra-firma vehicles (both two and four-wheel) and add the floating air transport IF it becomes available AND affordable in my lifetime.

But something tells me that the purist in me would want to have control over it as well. :)

Jeff: you can have your Red Barchetta, I'll take my gleaming metal (self-driving) air car. And that electronic parking brake on your CC sounds scary. Parking brakes are commonly entirely separate systems composed of drums and cables (completely manual) compared to brakes used for driving. Even if I had a self-driving car, I'd want a manual brake to make sure the sucker didn't take off without me!

I completely agree about the usage and context for texting as a communication mechanism...

I just think that there could/should be a better way for background communication from a MOVING vehicle.

My own subjective impression is that voice recognition technology has not evolved as quickly as many of us would like. It may be a tougher nut than many anticipated, even with legions of computing power.

Are there any usable "speech-to-text" converter websites online these days? If so, what is their business model?

I'm thinking about texting as in a communications mechanism. Seems like there are two reasons that it's become popular, one is that unlike voice talking on the cell phone it's asynchronous, but the other is that you can do it as a secondary process. Texting in school, texting as a semi-background process in other contexts.

It will be interesting to see how voice input evolves, it could very well be that my problem is that I've been conceptually poisoned by all of the systems which don't work terribly well. I suspect that in 10 years or so voice input will get to the point where it'll be able to do basic free-form input, and we'll completely bemoan the fact that kids those days have no grammar, and maybe then my perception will change. However, right now even advanced tools still more remind me of the hardware with a vocabulary of 32 pre-trained words that were touted as "the future" back in the '80s than they do of actual voice conversations.

I believe that texting simply arrived and was implemented first because it's relatively trivial to encode a keyboard compared to recognizing a person's voice.

Instead of texting, I truly believe that voice recognition is more of the end-game for a variety of reasons (some already mentioned).

I'd like to learn more about Audi's entertainment system implementation. How does it handle communication with other people?

Oh, on "The ultimate texting machine", Audi has been playing with an Android based entertainment system.

Just semi-related musing: I find it interesting that with vehicles like the Ford Focus and Toyota Prius becoming brands based on bringing the gadgetry to the lower end of the market, that at the higher end of the market the focus is primarily on higher performance.

I think the problem is you need to deploy a bunch of these things in traffic to remove the jams. See L.C. Davis (2004) Effect of adaptive cruise control systems on traffic flow which suggests that about 20% of vehicles in traffic using ACC (adaptive cruise countrol) can keep flow laminar no matter the density.

Interesting about preference for voice recognition, I think there'll be some fascinating sociology as that gets better, especially since the popularity of texting may be a reaction to social norms about taking on the phone in public.

I just think the Highway 1 example is a very poor use case and would likely not be representative usage of the technology? Just because they've done it doesn't make it compelling.

I would rather see a use case centered around a daily commute, with interactions with thousands of cars in heavy traffic. That seems a lot more germane and relevant?

Sorry Dan, but at a personal level I'll never find texting more enjoyable than driving--even in heavy traffic.

The notion of typing while driving just doesn't appeal to me. Of course, I've done it before, but I feel that voice recognition with the ability to "persist" messaging for distribution is preferable from both a UI perspective and from a human factors engineering perspective. This is not to say that texting doesn't have its place.

The Route 1 example is given in the article because they've done that: The car has driven itself from the Bay Area to LA (or vice-versa, or both) without driver intervention. Which is their point about texting: You're still looking at driving from the framework where texting is a distraction from driving. A generation or two from now, driving will be the distraction from texting; an option you might choose, but not the default.

As you mention, though, it's one thing to deploy a technology on a few hundred high end cars which will depreciate rapidly, it's quite another to build it on fleets of hundreds of thousands of vehicles whose owners will expect them to last for 20 years.

I think that for highly structured and repeatable routes (e.g. certain daily commutes to work) there can be a very high degree of relevance. But the choice of Highway Route 1 was an extremely poor example given in the article; I see close to ZERO relevance for that route.

As for texting in cars? I would MUCH prefer hands-free voice recognition technology which allows you to send "persistant" text messages to a given list of recipients. If we want to exchange information in real-time via voice we simply use existing hands-free cellular technology.

I'd much rather "speak" into my steering column than "text" into it. That way my eyes don't need to leave the road and I can enjoy the peripheral scenery if relevant.

As an aside, I absolutely hate the "electronic" parking brake on my VW CC. The electronics behind it have already failed once and it doesn't offer anywhere near the utility and control that a hand parking brake offers.

To a large extent we already do let computers drive the cars. They manage braking, shifting, follow-distances (there are adaptive cruise control systems which can manage stop-and-go traffic, which Larry would have appreciated), on trucks they've been assisting in lane-keeping for a decade now.

And, yes, there is a hacking community around such things, people "chip" cars constantly.

On public transit, I'm definitely not convinced. Buses can be fuel efficient if they're always full, but given the way we like to structure our communities you have to run them empty a lot to get people to use them during the commute times, so they're largely just an income redistribution mechanism. Carpools get the fuel efficiency much faster (although, admittedly, the real problem that public transit can solve is parking). It's likely that if we can get adaptive cruise control adoption rates to 20%, we'll remove traffic jams and dramatically increase road throughput. If we could get people to want to live in differently structured communities we could make public transit efficient, but my general guess is that telecommuting, online shopping and "the knowledge economy" will restructure how we view the commute and similar transportation long before energy costs get high enough that it's worth rebuilding our communities into denser cities, especially since we'll likely have a shrinking population before (or as) energy costs hit a point that would force such a restructuring.

On the enjoyability of driving: Yeah, people like that, and manual control of the vehicle isn't going away any time too soon, but I also think that's largely a cultural artifact, a remnant of an era that's over. The post-teens I know don't find driving as exciting as I did, if you want to do high performance driving really you're better off doing it on a track anyway, I think manual driving will go the way of individual aviation: Something that well-heeled individuals who can afford the additional insurance will participate in as a hobby.

I think that letting the computers drive the cars is asking for trouble in terms of the potential malware/hacker damage.

I agree that when a car is creeping along in traffic it is very difficult to pay the full attention that one needs, and that it's not a pleasant experience at all. Of course, the mass-transit advocate in me says, "If you are routinely travelling somewhere where you encounter those conditions, then you shouldn't be in a car at all, and now let us have the discussion about how America's transportation infrastructure priorities got so screwed up plzthx."

Jeff, in most cases I'll fully agree with you on this. However, my drive into work this morning on McGinnis Ferry Rd. would have been much more pleasant in the back seat texting as the traffic moved at 5mph when it was moving and was only moving twenty seconds out of each minute.

#Comment Re: made: 2011-05-12 11:05:58.36874+00 by: jeff [edit history]

For me personally, I'll always find driving to be exponentially more fun (in most instances) than texting. From the date of my first text until the day I leave my physical body.

To me there is a huge visceral difference between driving up the coastline on highway 101 and enjoying the passing scenery vs. texting while driving the same.