UAS News Podcast — 2012-05-23
Transcript
Subs by www.zeoranger.co.uk Hey everybody, welcome to the UAS News Podcast. This is your host Patrick Egan. We've got co-host Gene Robinson on today. Hello all.
Hey. In this week's installment, we're going to talk about some legal considerations for driverless cars and we have a special guest. Before we get into talking to the SUAS News Mr. Walker Smith about what he does and considerations for the cars, I kind of wanted to go over a little bit of the current events for this week.
Gene, do you have anything that you'd like to talk about? Yeah, there are some interesting press releases that have come out from the FAA concerning small unmanned aircraft. It's kind of interesting to note that the press releases don't necessarily match reality, don't you think?
Well, we've got to keep up with what we know from the FAA, right? That's great. Well, and the news stories are getting a little out of control too. I saw where Toscano got lambasted on Infowars and they said he was in full damage control.
And then, I don't know if you read that, but I got huffposed yesterday. I will say that somebody was a little loose and fast with the clothes. I sound a little nuttier than usual, but I think as far as I'm concerned, I think that the media is kind of turning on us a little bit. People are really up in arms about the privacy issues.
And one other tidbit I'm going to throw out here is I'm hearing from reliable sources that, get this Gene, the administration is telling the FAA to slow down because they don't want this privacy thing to work into a big issue. What do you think about that?
I am so shocked. They are politicizing an issue like this. I can't believe it. Just when you think it couldn't get any worse.
I used to call for, let's say that this integration effort needed a restart. Now I'm starting to think it might need a priest. If it gets a priest, it will totally die off. But I don't know.
It's getting kind of crazy out there. That's not really surprising considering the sort of politics that are going on. But there are a lot of agencies out there that honestly believe that they can start flying. And I think that the FAA is going to have a tough road to hoe when they try to start shutting these guys down after their 90-day deadline has already passed.
Don't you think? I think so. I've been talking to people out there. A lot of these companies, their phones are ringing off the hook.
People are calling them. They're interested in buying. What are we going to do?
How are we going to do it? I think some people are saying full steam ahead, damn the torpedoes. It should be pretty interesting in the not too distant future. All right.
Well, let's move on. Now that we've uplifted everyone's spirits, I'd like to welcome Brian Walker-Smith. He's a lawyer, former transportation engineer, and current fellow at two Stanford centers, Center for Internet and Society and Center for Automotive Research. And he researches law and policy of self-driving vehicles.
Brian, it sounds like you're a busy guy. Well, Patrick, I'm thrilled to be here. That's good. So yeah, do you have a social life or are you just a fellow at these two places at Stanford?
You're busy all the time doing that. What's going on?
So one of the offices where I work, five feet from me, is a self-driving car. So it's a fun place to hang out. And it's not one that I can complain about spending time at. Are you working in the Vale Center at Stanford?
Exactly. I'm over at Vale. I'm also at the law school and over at one of the other engineering buildings. So I always joke when I want to...
pardon? Do you drive or do you ride a bike or driverless vehicle between the two offices?
How does that work? I bike and haven't made that self-driving yet. Well, I did enjoy the Vale Center at Stanford. It's pretty nice.
I got a little chuckle out of the SOPs for operating the espresso machine there. That was pretty funny. But I'm not going to lie. You're also going to be...
we met at the Silicon Valley Robotics Block Party in March. And I think that's where the first time we kind of conversed. You're also going to be a guest at the Silicon Valley Chapter meeting where you're going to come and talk to us, which is a good thing. And if anyone's in the area near Moffett Field, NASA Ames, you can come by on Thursday, next Thursday between 6 and 8 and come on in and meet with us there.
But I wanted you to introduce yourself, Bryant, maybe talk a little bit about how you got involved in this, why you got involved with it, a little background if you could. I'd be happy to. So as you mentioned, I used to work as a transportation engineer. And in that role, I really confronted this American ideal of cars as the ultimate expression of and tool of freedom.
And that's the idea that we have of a car, right? You get in your car and you can drive anywhere. The road's the limit. But that was frustrating because cars don't necessarily mean freedom.
There are issues of congestion. There are mobility concerns for people who are disabled or young or elderly. You have 10,000 people turning 65 a day right now. And to the extent that we are dependent on cars, those are populations that are ill-served.
And that ultimately led me to look at transportation much more holistically and to look at transportation as a policy issue and a legal issue as much of as an engineering issue. That led into law where my particular focus was administrative law. That's legislation and regulations and decisions that are made by state and federal administrative agencies like, for example, the Department of Transportation or you were alluding to the Federal Communications Commission or the Federal Aviation Administration, agencies like that.
That's been a real useful background for looking at self-driving vehicles because so much of this is happening at the regulatory level and at the standards level. I also focused in international law which surprised me to learn actually relates to self-driving vehicles. So we say that in law we deal with precedent. That is we look to what's already been decided.
We look to how the law has already dealt with issues. And when we're dealing with something this new like self-driving vehicles, there is no precedent. There is no cases or law directly on point which means everything is relevant. So as you said, I stay busy.
Yeah, it sounds like it. You know, the Schlemacher gets a little deep in the federal realm and you know, like the Department of Transportation, whatever. Me and Gene have a little bit of you know, experience with these guys. And I know that it can be trying when you get at these types of levels.
You know, you hit on some other good points, you know, about the freedom in Americans thinking about cars and everything else. And you know, I got to be honest, I don't have the love affair with cars that I used to. I mean, you look at these television car commercials and they're you know, driving up Highway 1 and it's totally empty and the guy is driving the road and you're like, yeah, last time I was out there, I'm behind a convoy of Winnebago's and I'm not enjoying myself. I'm getting carsick and I'm yelling and screaming out the window.
You know, I guess there was a good Arthur Schopenhauer quote about buying books and it would be great if you could buy the time to read the book. I think it would be great with a car if you could buy the road and you could be out there by yourself. Do you see in any of your research, do you see this kind of changing?
Is it like people aren't really just going for rides anymore? It's kind of a utilitarian type of thing. I just need to get from point A to point B or what are you seeing out there as far as the driving demographic goes?
So you're absolutely right that it is a huge shift from 100 years ago when engineers actually built, they were called Park and Pleasure drives for the sole purpose of treating driving as the park experience, as the recreation in itself. Clearly the vast majority of transportation is not that today. On the other hand, a lot of the transportation is not perhaps strictly necessary. That is, it is incidental.
It allows people to do something else but if driving were not available, they might take other options, all the alternatives. For example, they might shop closer, they might drive more locally, they might simply use the Internet for services that would not require the use of a car. In that sense, we are seeing a demographic change. There's been a good deal of research that's looked at younger generations who are much less interested in getting a driver's license or owning a car, either as a means to their independence or as simply a status symbol.
We're also seeing probably the effects of the recession on just general vehicle miles traveled which has been for the last century a huge increase in total miles traveled to 3 trillion miles in the US last year. That's leveling off. It will be very interesting to see if that trend continues or not. Brian, wouldn't you think that there would be a regional division there as well?
There would be more interest in self-driving car in the urban areas versus the rural areas or the ranch areas where there's wide open spaces? It's certainly possible. One of the questions is when we see the end goal of this technology deployed, that is a fully capable self-driving car that really doesn't require a driver in any sense. That's a ways off.
Who will that actually appeal to? It could certainly change the urban environment, anything from parking to circulation. It could also change rural land use that people may be willing to live further from their jobs if they can get in the car and catch up on their sleep or if they can bill an extra hour of time while in the car. You ask who would it appeal to.
I've got to tell you, I think there are several scenarios where it would be something that people would want. Maybe an emergency situation, I'm having a medical condition and I can hit the red cross button on the dashboard and it takes me to the emergency room. Or you hit the breathalyzer and either the car can make the determination that I'm drinky poo or we're heading by the corkscrew for a nightcap. I think in either one of those situations, at least I'm not getting the DUI, I've got driver service, it's a good thing.
I guess that would maybe lead into another question that would be, is there a pushback? We know from the drone, as they call them, an unmanned aircraft system thing, people are a little apprehensive about this technology. How about the self-driving vehicles?
Do people have apprehension that you've talked to? We tend to love technology until it's actually real and feasible. Right now, for all the public discussion of self-driving vehicles and all the research on them, we don't actually have a vehicle that is capable of driving itself without human oversight yet. When we're talking about these spectrum of technologies, it's an open question how the public will ultimately react to the kind of technologies that are possible, the kind that actually show up in production vehicles and then the possibility of a truly self-driving vehicle.
Again, you see a range of reactions. A lot of people are enthusiastic, a lot of people are apprehensive and some just say, well, I wouldn't be willing to spend a lot of money on that, that I trust myself as a driver and I don't trust the other people out there, but I would not pay a lot for that. The interesting thing will be, when this is an actual option, will people see the benefits in the way that they have for smartphones?
Where if you look back 10, 15 years, there was also less apprehension but more just general indifference. Why would I need a phone that does that?
Now, I can't live without it for a hundred bucks a month. I guess another one would be because you're talking about this eventuality. I'm in Napa this weekend up in the mountains and I'm buying some oil for the truck and talking to the guy behind the counter there and the guy at Napa says, well, we got talking about this and he says, when this becomes an eventuality, where does the responsibility or ultimate liability lay for mishaps?
Is it going to be the driver or I guess the owner or the guy who's writing the code or maybe you can enlighten us on that? Absolutely. Liability often comes up as a concern for manufacturers and others in the chain of supply and for insurers as well. The first thing we can say is that the hope is that automation will reduce the crash rate.
Now, that's not certain, but that's certainly the hope of researchers. On the other hand, it's not likely to eliminate all crashes, so there still will be some. For those crashes that still occur, as you suggested, manufacturers will be theoretically liable for a greater share of the crash damages than they are today. I say theoretical because law operates in strange ways.
We don't ultimately know what judges and juries will do or how strong a case defense lawyers or plaintiff's attorneys will have. Now, if we think about it in the abstract, it's not a difficult problem. That simply means that costs that are currently borne by drivers in the form of insurance are in the future borne by drivers in the form of a higher initial retail price or a higher lease price or a higher charge for a subscription self-driving service. The real obstacles are twofold,
and that is getting to that future because there's a lot of uncertainty both about the extent of that liability and the extent of the reputational risks that an established company would take on by offering a vehicle that may have an uncertain public reception.
It's the difficulty of pricing that uncertainty that may be more of a barrier than the actual functioning of the liability system. There are certain ways that we can look at providing greater certainty. One of those is the collection and use of data. The other is looking at how recalls are handled.
If a vehicle is put out there that needs a software update, should that require you to bring the car in? Should the car drive itself in?
Should the entire autonomous system be shut down until the manufacturer can design a new software patch? Should it just be updated remotely in a beta version?
These are open questions. They're all really good questions. I was thinking about some of those. We tried to relate it.
Me and Gene are usually relating things in a 3D environment where if you have any of these issues that come falling out of the sky, and people have obvious apprehensions, and we've advocated for cracking the door open with small, frangible aircraft. Same thing with a car. To me, again, this seems like a no-brainer. There's still going to be issues and mechanical failures and whatnot.
When I look at it, I think you're going to have something where you're going to have a safer vehicle with reduced crashes. Are the crashes going to totally go away or mishaps going to go away?
I would say no, they're not. Again, let's look at the statistics, the DOT statistics for cars with drivers in them. I don't know about you guys, but I blow down the road here and you've got these people on the cell phone or they're texting. I don't want to mention any names or putting on their makeup or whatever else.
You think you could drive the car? Moving forward, it would be great if you could hit the autopilot and you could put on your makeup on your way to work. I don't have to risk my life out there. That may be something as far as a takeaway you could consider.
I still think there'll be mishaps, but another question that feeds into it is what kind of software-based systems are these? I used to make jokes about with the unmanned aircraft that I hope these aren't Windows-based systems because we could have problems. You're absolutely right. If you look back to that era, there are all sorts of jokes on the Internet about a Microsoft car or an Apple car, what it would or would not do.
I'm actually amazed by the particular software platforms that show up in certain components. It's a surprising range of software. Ultimately, the production versions of any of these systems you see will have extensive testing. One of the open issues is what standards should be established for that testing?
When you're dealing with such low rates of incident, how you provide certainty that these systems are in fact better than humans? If you were to, with statistical confidence, say that a vehicle was safer than a human driver, you would have to drive at about 750,000 representative miles without incident. That's a pretty big number for research purposes. Humans, like you said, are very imperfect drivers.
Some 90 to 95 percent of all crashes are caused at least in part by human error. On the other hand, we're incredibly good at driving. The range of circumstances that we do confront are truly astounding. Of course, a little bit at a time, I think the public is being exposed to portions of the autonomy that the vehicles are coming up with.
For example, the Japanese have a car that will park itself. That small increment is being introduced to the public so it gains acceptance. Wouldn't you think that that's the way it's going to be eased into the public view?
You are absolutely right that these technologies fall along a spectrum. Self-parking cars are one example. The self-park systems that are available on the market from Japanese and other automakers will generally park themselves as long as you apply the brake and gas. The human is still in some sense in control.
There are other technologies that also represent some driver assistance. You think of electronic stability control or anti-lock brakes or drive-by-wire steering and braking. Those you simply suggest to the car that you want to turn or brake. The car says, well, let me see.
You know, I think I can do that. Let me do it for you so that you don't spin out of control. Yeah. Oh, sorry, Gene?
No, go ahead. I'm sorry. I stepped on you there. Go for it.
Continue. Yeah. We're seeing that incremental technology ending up in production systems. We may ultimately see deployment scenarios for self-driving cars that reflect that.
That is advanced driver assistance or systems that are restricted to particular geographic areas. Here I'm thinking of military bases or retirement communities. We might see the technology deployed only on certain freeway corridors or under certain conditions. You think of US 101 and how many people would pay for an extra hour of billable time on that road every day.
We might see them in mixed traffic or we might see them in dedicated lanes. You know, they may have their use as an LA service. You can just get out of your car and it will park itself in a controlled environment or some limited deployment of taxi. It's a mixture of what technology is available and what are the reasonable commercially prudent scenarios to use it.
What about, you know, I like kind of how you're into the moniker of self-driving vehicles which kind of opens it up a little bit. I mean, what about, you know, farm equipment or installations you mentioned, military, but what about maybe irrigation districts or whatever else and vehicles that could do jobs on their own, you know, maybe like water trucks or whatever else or harvest weed or whatever. Is that a consideration also or is that outside your purview?
Absolutely. Those raise different and in some ways more limited legal issues, but we already see considerable progress in autonomy in agriculture. So, vehicles that farmers can buy today that while not entirely autonomous are pretty far along, certainly further along than some automotive systems in the area of mining. Rio Tinto which is a huge mine operator is completely automating a number of its mines.
Anything from the huge trucks that drive down to the trains that take the materials to be processed. Well, I think that, you know, you just touched on trains too. I mean, there are a lot of systems around the world, you know, where there are no drivers in the trains and they are all automated. I think this is just kind of like a progression and it's kind of funny when you talk to people about this technology and the automatic cars and everything else, you know, even you're saying, oh, we'd have to drive 750,000 miles or whatever.
You know, we're used to the DOT, arm of the DOT being the FAA, you know, telling us we have to have numbers of like 10 to the minus 9. So, those are some like astronomical numbers. I know I was telling them, you know, I think somebody calc'd it out that it would take 200 years for me to fly to get that kind of testing and I was taking vitamins and what not and I just didn't think I was going to fulfill that test, you know, matrix. But, you know, so it's not too bad with the car.
But, you know, what do you think about, I know that Google, you know, has jumped in this with both feet and some people had kind of said, you know, well, you know, the party's over, Google's in, there's no way out, there's going to be any room for anyone else in this market. What do you think about that?
Google is one of many companies that is researching this. I would say Google deserves a lot of credit for fostering discussion in public and for pushing regulators to begin really focusing on self-driving vehicles as a distinct possible technology. But, you talk to a number of automakers and suppliers and then they have their own research going back decades. So I don't think I would say that the party is over and in fact I would say that any problem, particularly that of road safety and mobility, you really need smart minds from different perspectives.
And that means a lot of things. For example, you look at software as generally being developed in the order of, say, six months versus cars being developed in the order of every six years. Traditionally, cars and automotive transportation have been thought of as a product rather than a service. And that could change.
So instead of buying a self-driving package that simply starts the moment you turn your car and lasts until you junk it in the junkyard, it could conceivably be part of a subscription package that's only available at certain times on certain routes and that needs to be serviced every now and then to continue using it. And it also goes to the idea of our shared digital infrastructure. So Google is brilliant at collecting and processing and understanding data. These cars will require vast amounts of data and produce vast amounts of data.
Just one example are the highly sophisticated maps of roads that, for example, Google's system relies on. So the route on which it drives must be carefully mapped out ahead of time. Will Google collect those road data?
Will they become a public service? Will other companies collect and then sell those road data, including data on say construction, crashes and congestion, all of which may be necessary for completely autonomous operation?
Definitely some new business opportunities were just brought to light there. You know, I kind of have the same feeling about Google jumping in. I know that Google had expressed some interest in unmanned aircraft and that was right about the time that the FAA kind of clamped down on it, which was too bad because I agree with you, you know, there's some smart people over there, there's a good business model, it'd be nice to have some of that jumping in on our behalf.
I think the FAA though, that somebody must have looked at that and went, whoa, you know, that's too crazy, we can't even deal with those people and jumped out because they're not really, they're not, I mean the FAA thing, I had another conversation with some people today and they were like, you know, where are the engineers, how come they don't have any engineers over there?
And I'm like, whoa, you know, let's not start bringing science into the FAA equation because they won't know what to do with it. But anyway, I don't want to get too far off the track with that one. The conversation today has been very enlightening. I really enjoyed it.
You know, again, it's kind of a different mindset for Gene and I thinking about a 3D environment and being in a 2D environment. Well, like you can say, you're playing to a little more space maneuver but it's harder to pull over. Exactly. A little bit.
It's a little bit more difficult if you have a little issue there. But I look forward to seeing this technology in the future and I look forward to seeing you at the meeting. I also wanted to say that Brian is going to be at the AUVSI Driverless Car Summit, June 12th through 13th in Detroit. And you are going to be a panelist discussing the topic of understanding the legal issues of driverless car integration.
And I know that AUVSI is really putting that one out there and I think that show is going to be very big. I'd like to thank you for being on today's show and hopefully we'll maybe get to talk to you again in the future. Thank you both. All right.
Have a good day and until next week, we'll see you then.