Smart Driving Cars episode 407- New Automated Driving Regs

It's the Smart Driving Cars podcast. We are glad you're tuning in. I'm Fred Fishkin along with the Faculty Chair of Autonomous Vehicle Engineering at Princeton University, Alan Kornhouser. Hi, Alan.

Hey, good morning, Fred. And with us from Sweden, consultant and co-author of the real case for driverless mobility, Michael Senna. Hi, Michael.

Hi, Fred. Hi, Alan. Great to be with you again.

We have a great discussion on TAP here, I think, and Alan, the Department of Transportation, put out a notice and request for comment on a proposal for a new United Nations Global Technical Regulation on Automated Driving Systems.

NHTSA under the DOT is seeking public comment on the draft to help inform the U.S. government's position on this. You've answered the call, Alan, sending in your comments. Give us your thoughts. Do we get it?

Yes. Yes, I answered the call. Since they asked, I figured I might as well put something in there because the real challenge here that we're facing to try to get some good out of this technology that we've invested in for who knows how long.

I've been doing trying to do automation and transportation to improve mobility in cities for over 50 years. So it's been, and how successful have I been? I guess zero. So, you know, I don't know that that means anything, but it means something to me.

And really, the question seems to be, besides economics, which it has been, is how much risk are we willing to undertake for what reward are we going to get out of it?

And of course, in my, I guess, venture into this space, I think I've focused on the reward side and assumed that, you know, we'll get the risk side in shape such that we can take advantage of this thing.

And more recently in the last 20 years, with respect to having machines do the driving, it came obvious to me at least more than 20 years ago that, my goodness, this is the way to go because the opportunity to really go through the work to make it work.

You could do it with just one vehicle. You didn't have to invest in the whole system. And if you've got that vehicle to work, the nice thing about if that machine works, then you just replicate the machine.

And then you have two. If your replication is easy, then you have four, then you have eight and so on. You don't have to start out with a lot.

And you can really invest your resources in getting that one to work. And in some senses, this is really what we've been doing over the last 20 years is just getting a vehicle to work. Why?

Because the fundamental aspect of this approach is to share enormous investments that have been made in the past and just to go and use them.

And some use them for free. Heck, if you're only going to have one vehicle out there, my goodness. And that's the support system, the way system, the roads that have been built for and found to be a value for another other machines, other things.

And we could go with this new machine and just use it. And of course, you know, what has been the underpinning of this 20 plus year process has been.

We don't need anything else except what we already have in that infrastructure. And the thing that we've all been trying to design is something that can just use it.

Go in there just want to be a very marginal change, not affected a whole lot. They grab the value out of that thing. And then with two and then with four. That's why, you know, don't want a bunch of new communications, don't want a bunch of new things, new hardware, new rails, new who knows what.

Ah, you know, how about just some paint? How about just a smooth surface? But my goodness, that's the system, the vehicles, the old machines or current machines shouldn't come all machines can just go use them.

You know, I mean, those the paint and the smoothness, Jesus, they want that too. So not asking for anything else.

Because the previous views on how to do automation, how to do a machine, move people.

You had to build the infrastructure, even for the first machine.

And that infrastructure wasn't just small thing. If it was going to do anything, it had to be big.

And so of course, then who's going to pay for that? You know, I think in 1975, my class and I, you know, designed a statewide PRT system for New Jersey, 10,000 stations, 10,000 miles of guideway.

We went down New Jersey's DOT, who look at the great mobility we could provide, who to provide it, we needed the structure, the infrastructure.

I don't think they even bothered to laugh. Okay. Get out of here, kid.

You know, I mean, it was, and of course, because, you know, to do that, you would have to go to Wall Street and say, as for, I don't know, quarter of a trillion.

And not only that, you would have to, you would have to go to how many public meetings so that you could put your guideway up next to people's bedroom windows.

I mean, I mean, talk about being naive.

Well, you know, I just come out of wanting to go to Mars. I thought, whatever. I mean, it's, they didn't even laugh.

I mean, they just walked out and deservedly so.

And it wasn't until 20 years ago, it took 30 years, almost 30 years from there to get to a point to say,

I mean, if you really want to provide mobility, if you really want to take advantage of machines to do the mobility piece,

you had to use the existing infrastructure. You had to start. You didn't have a chance.

It took me at least 30 years to learn that lesson.

That's where the whole, at least to me, even before the DARPA challenges came to me that Jesus,

let's just put the intelligence in the vehicle as opposed to the guideway and let the vehicle deal with whatever it needs to do to not crash.

As opposed to putting guardrails around everything and bumpers and things to make sure it doesn't run off.

And this is how we've gotten here. And we're here. And my goodness, we're so close.

But the discussion is now still about, the discussion is still about the safety thing.

And the unfortunate part of the safety thing is that it gets embroiled into the existing technology, the existing machines.

And if you really look at what is the safety problem, at least I think when I really look at the safety problem,

it's the thing that the problem is the person that we're trying to take out to do the automation around.

The human in the loop is the challenge. And of course, you can't take the human out of the loop unless the human wants to take him or herself out of the loop.

And it's their system. Can't. And you can't say, Hey, behave out there. If you behave, this is trivial problem.

It works if everybody behaves out there on this thing that's for them.

And of course, the tough part of this thing is who dealing with the misbehavior.

And, you know, two ways to deal with misbehavior, stop the misbehavior or somehow avoid it.

It's impossible to, I mean, people have pulled their hair out trying to stop the misbehavior.

It's not that they haven't tried. It's not that they haven't developed, you know, systems that limit one from speeding too much.

It's not that they don't have breathalysers that tell you and tell somebody, Hey, you're too drunk to do this.

It's not that, you know, as I like to say, Apple even says, you know, don't use. I think you're driving.

Unfortunately, Apple lets me lie and tell it. No, I'm not. I'm not. I'm not.

And to get Apple to stop, you know, setting me up to lie so that I can do it, I mean, it's essentially impossible.

Why? Because then people won't buy it. People don't buy it or buy another one or let them.

So the problem moves from Apple to somebody else.

So we're stuck.

The way we've been dealing with this, the way Nits has been pulling its hair out over the years dealing with this,

I don't know. I don't think they've given up in collision avoidance.

They've focused on collision mitigation.

Rightfully so.

If it's going to happen, then let's make sure that whatever happens is less impactful to the things that we cherish.

So, you know, anti-lock brakes, seat belts, airbags, crush zones in car, you know, really good work, fantastic work.

But through that, through doing that, the crash mitigation, it's made it acceptable to us.

Because of the whatever gazillion crashes that occur today, we're only killing this many.

If we didn't have the seat belts, if we didn't have the airbags, if we didn't have the anti-lock brakes, if we didn't have it, I mean, think of beating in the 55 Chevy today.

I don't know.

That steering wheel goes right through your chest.

The cigarette lighter goes into your eye.

So there's been a lot of beautiful work done.

And can we do more?

Well, heck, I think we can do more by just the automation, the things that we're trying to do.

And it should be appreciated that the fact that if this machine can see better than I can see, can react faster than I can see, cannot blink, cannot do those things, we should be all in, at least to see what good it might do us.

That should be the discussion.

To then say, no, no, no, we're not going to look at any good until we have no collision.

I don't know.

I quit.

I can't do, you know, I spent 30 years, whatever, trying to whatever, I don't, I don't, I won't have 30 years.

Ain't going to happen.

And it's what a shame.

And I think, you know, what's great about what what's happening now is I think we can, we can begin to have that discussion.

Through the beginnings of what is being set up here with respect to the comment.

And so, you know, that's really where where I tried to take my testimony to say two things we've got to do. Hey, of course we wanted everything, but we want to look at, don't put the hurdle to be crash avoidance.

Because, you know, sure, maybe let's just have the hurdle be crash mitigation, because that can add to it.

My goodness, seat belts, we wouldn't go back from that airbags, we wouldn't go back from that in a lot of breaks, we wouldn't go back from that.

Being able to see better, being able to react better, we should be praising those things.

We'd love to do more.

Don't get me wrong.

Now, within that framework, let's look at what we get out of it.

What this now gives us in addition, because this is going to give us more.

I don't know.

Maybe it's unfair to say the fact seat belts, they just did crash mitigation.

Certainly in the beginning, they were extremely uncomfortable.

And whoever figured out that if they put slack on this thing and allow me to move so it's not choking me, I'll put up with it.

I mean, I don't, I don't, I think probably 30 years ago, I stood up in class and I said, I'm never going to wear a seat belt.

I mean, those things were and all the good work that was done to make them palatable.

Praise that.

So anyway, so it's in this context.

Not to get here, you know, people have been working diligently to try to come up with with requirements, frameworks and so on.

And stuff that stuff that I can't do.

But we have Michael Sene here.

Michael, Michael and I are different.

I love differences.

Michael

was part of the heavy lifting that got us to this point so that we can have this discussion.

And I think, I think, you know, I think we deserve to discuss a little bit the heavy lifting that went on, you know, to get to this point so that we can begin to have this discussion so that we can then put rules out there so we can take advantage of the crash.

mitigation opportunities of this and maybe the crash avoidance.

If we get some of that too, of course, we want some of that.

Okay.

So that we can then

harness the value proposition of what this thing does for us, which I think Michael and you do Fred think is just enormous that we can't do in any other way.

If we could, if we could do it in the other way, we would be doing it.

Okay.

But we're not doing so Michael jump in here and sorry for the long introduction.

You know, I just, you know, it's important to to provide the introduction that we were not just starting.

We're not just starting with, well, we have, we want to have driverless cars or autonomous cars, whatever they're, whatever we're going to call them today.

And that they're going to solve all the problems with people having accidents and people driving drunk because if you remove the person, then of course you're not going to have the bad, all of the bad things that people do.

But as we found during these 20 years that solving the technical problem of getting robots to do what humans are doing is today and have been doing for the last 100 years, which is getting in a car and driving it and ending up where they wanted to go without killing themselves,

injuring themselves or injuring other people or killing other people, whether they're in other cars or people or by the side of the road.

And we've been doing this now for quite a long time.

It's not the easiest thing to do.

I compare driving to a game.

Take a friend of yours from, let's say from Nairobi.

I've never seen a baseball game before.

Take them to Yankee Stadium and sit down next to them.

And what are they going to get out of this game?

They'd have no idea what the rules are, why people are doing what they're doing, whether the people on the field or the people behind the plates or close to those things called bags.

And you try to explain it to them as the game is going on.

It's a very, very difficult process.

Driving is exactly the same thing.

It's a huge game.

The only big problem with driving is that it's played in an arena that's as large as the planet.

And the rules that are required change constantly and unfortunately many people don't abide by the rules.

And the conditions in which this game is being played are changing by the minute.

It's raining, it's snowing, it's dark, it's light, it's sunny, all of these things are occurring.

And you're trying to guide your tool, your weapon, whatever you want to call it.

But you're involved in this game of getting from where you are starting this game in process and getting to someplace else.

And you win most of the time.

The largest number of people actually do get to someplace without causing any damage at all to themselves or to other people.

Now we're going to impart that kind of knowledge that's required to do this to robots.

And it starts with giving them a basic understanding of what the rules are and then giving them the equipment,

designing whatever they need to have as their tools to be able to manage this vehicle in the arena.

The arena being the roads, wherever they are at whatever time.

So it's a combination of having the rules by which the people can operate.

It's one thing to say, you know, here's a robot, you're going to play chess.

Now here are the basic rules, go off and do that.

Or you're going to run a hundred yard dash.

You've got four rules that you have to abide by.

The first one is don't go over the line.

The second one is don't start before the gun goes off.

Stay in your lane and keep running past the finish line.

But now the rules for driving are extremely complex.

There's an organization that was set up to try to put rules into language that can be interpreted

by the people who are going to design the tools that we use to participate in that game.

It's called the WP29 World Forum for Harmonization of Vehicle Regulations.

It's this group who have had the responsibility for developing most of the requirements

for the way that our systems that have been developed over the last 50 or 75 years have been developed.

And these rules and regulations have gone into the sets of requirements

that manufacturers use to build their systems and the tests that are used

to make sure that those systems and vehicles are built in ways that make them capable of being able to be driven.

This organization has taken on upon themselves over the years to develop new requirements.

And during the last several years they have been working very hard to develop requirements

for vehicles that will be driven by the vehicle itself, the robot inside the vehicle.

It's easier to think of it as a robot than to try to think of it as some sort of magic wand that guides the thing.

And it's this group that has developed the general requirements that NHTSA has now decided

they will incorporate into the Federal Motor Vehicle Safety Standards,

which is what in the United States, which is what are used by the automobile companies

to build their cars and build the systems in their cars,

and then to test those systems to ensure that they are drivable.

And NHTSA in the United States, other organizations in Europe,

we've got something going on similarly here in Europe with other organizations with the European Union,

but in the United States, NHTSA has asked for comments on the work that's been done

for moving the ball down the field for driverless mobility.

And this is what Alan is referring to, that those comments that are coming in will help NHTSA to decide,

will we use these, do we need further work,

is it more important that we get something in there now so that we can begin to guide the process in a better way?

And we're there right now, and hopefully I've seen the requirements and looked at the requirements

and working on other types of requirements with the groups in the WP29.

But this piece of work that's being done right now is very important to be able to go to the next step.

Explain to people if we can. I mean, there are robotexies operating in limited fashion, I suppose, in the United States.

In China, I believe, and I'm not sure how many other countries have them operating at all.

But if this is taking place now, tell us what needs to happen and why it's important for the United States

to be on board with what the UN is formulating here.

The United States and NHTSA is part of the signatures to the 1998 agreement,

so that even though it's not using the same process of testing that's done in Europe, which is called type approval,

so you somebody test the car and then it's approved for all of the European countries,

and it's done in Japan and several other countries as well.

In the United States, you have the Federal Motor Vehicle Safety Standards, which are very similar.

The requirements of type approval and FMVSS are very, very similar,

but in the case of the United States, each car company does its own testing, so it's called self-certification.

But essentially, it's the same process, self-certification versus certification by a testing agency.

There are pluses and minuses on each side, but essentially, it's exactly the same thing.

So the United States does take currently, does take the requirements that are done by WP29,

but in Europe with EU, it's an automatic process.

They don't go through the same steps of having to have the review and decide,

are we going to do this directly or are we going to make some changes.

When it's done in the EU with WP29, once it's decided on, then it goes into a regulation

that's then put into the rules and regulations of each country,

every country that's part of the type approval process.

So that's the main difference between how it's done in the US versus in China has its own process.

China uses its own process.

It's not the same as in the US and it's not WP29 or type approval.

It's important, why is it important?

Car companies are making their cars, have been making their cars globally since the beginning.

Up until the 50s or 60s, most of the cars sold in Sweden were made in the United States.

But there started tariffs and various things after World War II and unfortunately,

it got to the point where there were separate markets and they developed separately

and you've got metric systems and inch systems and all kinds of different regulations.

It would be great if it was one system, but I don't think at this point in time

I think it's better to try to coordinate rather than try to force everybody into one set of regulations.

But it is possible through WP29 to have a set of regulations which then can be incorporated

into the different regimes that exist in the US, Canada, China versus the type approval countries.

Interesting thing is both the authors of The Real Case for Driverless Mobility

and in Alan's comments here, it seems like Alan's trying to focus what they're doing,

at least in his comments to NHTSA here, trying to focus things beyond the safety

that you should be looking at the societal value.

To me, I think that's sort of hopefully the important part of the comment that I'm making.

I think that the work that's been done to this point is fabulous.

By very good people spending an enormous amount of time, people knowledgeable about this,

from my outside view, essentially zero political influence on the darn thing,

just trying to deal with the technical problems associated with trying to make these things,

of course, as safe as possible, because in a sense the public oversight,

its number one responsibility is to protect the public and therefore the safety oversight is exceedingly important.

Because you can't have a bunch of loose cannons running around on the deck of the ship.

I mean, everybody knows that from day one.

So in some sense, you need to bring a bunch of adults into the room and come in and put together

some good solid technical recommendations on how to do this.

As an oversight, one should also look to see what it is that one might be able to do with this,

the value proposition, because in the end, the risk is going to be traded off against the reward,

and it's either going to happen or not happen, and it's going to happen only if the reward ends up being greater in the risk.

Michael's time is limited here, so let's go.

My last point here is the organizations that are putting regulations into effect,

whether it's in the United States, in Europe, wherever it is, they're not concerned with,

I'll make it very direct, very simple, they're not concerned with the business case.

They don't care whether you're going to use your car to take your kids to school

or to take yourself to work or go to the gym.

That's not important for them.

What's important is that the car, whatever it's used for is safe and does what it needs to do safely.

The problem that we have with introducing the concept of someone, a robot driving the car,

is that robot has to be able to drive the car at least as well as the human being, at least as well.

Making claims that it's better, that's not really important.

They have to do it at least as well as the person who's driving the car.

The real reason for doing this is not to make cars safer, because making cars safer,

as we said over and over again, we can eliminate 60-70% of the accidents and make them safer

if people just do the right thing.

They don't drive drunk, they don't speed and all of these things.

We could put things on cars to make sure that people don't do that.

But the real point of having someone, a machine driving the car instead of a person,

is that the machine can drive itself without having the cost of the person included

in delivering the ride.

If you say, I'm going to drive my car anyway, I'm going to take myself to work,

then you can look at it in another way and say you can do other things while you're getting yourself to work.

You've got benefits there, they're economic benefits, and that's used.

But the big argument that's used for driverless mobility is that it's going to be a lot safer.

That isn't getting us anywhere.

The problem we have is that's not a big enough reason to be able to put all of these vehicles on the road.

And as Alan has said, we can use the safety aspect, we can use the increased safety aspect

of further mitigation of accidents, of reducing the number of possibilities of cars

that getting involved in accidents that are being drive by drunk sentient and being in speeding

by having the regulations built into the vehicles so that they don't exceed the speed limit.

And it's easier to do that if you've got a machine than with the individuals

because we know we've been trying this for 50, 60, 70 years to get individuals to stop speeding

and to stop driving drunk and all that other stuff, use your seat belts, etc.

So the positive aspects of this, both from the standpoint of safety as well as from the standpoint

of the social benefits that can accrue from driverless mobility are extraordinarily high.

I'm not going to say enormous, they're just extraordinarily high.

And that the benefits can come, will come, but they have to be coming both by assigning the rules

for how these systems are going to work as well as building the systems

so that the systems can actually do what they're supposed to be doing,

which is driving a car at least as well as a human being.

The website for more is michaelelsenna.com.

Thank you so much for spending time with us, Michael.

It's really great having Michael on and I'd like to expand on what he just said.

At least some of us believe that or envision that with this machine operating a car

allows one to deliver to society enormous benefits that it hasn't been able to deliver to society

with the existing transportation systems that exist on the planet today.

This is new, allowing us to do a better job of mobility for more people and it becomes so obvious.

I mean, it's just ugly in some sense what our current stable of mobility systems does

in terms of just not addressing mobility challenges of people who deserve to have good mobility.

The only reason we go from A to B is to improve our lives, otherwise we stay home.

Nobody's stupid out there from their own personal desires and they deserve to have those desires.

It may not be the same desires I have, but my goodness, diversity is good.

Trust me, you wouldn't all want to be like me.

I wouldn't want all of you to be like me.

I didn't tell my class, I'm going to teach you something.

I don't want you to learn it all the same as I do, aren't it?

I mean, there's the diversity, the different interpretations that each of you are going to put on

with respect to what it does for you.

That's the beauty of the human species.

If we were all the same, maybe we're too similar.

We don't want to go too far down that road, but it's true.

I mean, there is not a level playing field in mobility.

My goodness, I have to fly commercial.

I mean, I can't imagine being a chauffeur and not too bad flying commercial.

But in other places, it's like, whoa.

And yes, the car, the conventional car has been made available and things have been done.

And people have worked really hard and they've leaned things one way and another so that a lot of people,

well, essentially, other than walking, 95% of the person trips in the United States are taken in a car.

But I think mass transit is too, it's probably Uber, Lyft, whatever it is, 0.2 bicycles.

I don't know, doubt they're 0.2, but somebody could correct me.

Everything else is by car.

And of course, if you have one that you can drive, then the level of service is like phenomenal.

Right outside the kitchen door, blah, blah, boom, whenever you wanted to do that and probably the most important thing about that car is

why it's so popular is because you have a parking space where it can then sit there and wait for you until you're ready to come home.

So you put this in any mode split model.

That sucker says, if you have it, use it as it should because that's what human behavior is and has been for quite a long time.

And if you don't have the, if you're not permitted to drive because it requires a permit, society says, you got to be good enough.

Boy, you're not good enough.

You're not old enough.

You're too old.

You're too. Who knows what?

They say no.

All of a sudden, this fantastic mobility opportunity, if you can afford it because it's not cheap, it'd be nice if it was cheap.

Whoa.

Paratransit service, which is on demand responsive order it, I don't know, at least the day in advance time window, who got to get a big one.

So it's not really that demand responsive.

I think New Jersey Transit 2023 spend $120 million to give 1.3 rides.

You do the arithmetic 90 bucks of ride.

And it's not that people at New Jersey Transit work hard. They're smart.

They try like heck.

And with all that effort to try to give some folks who don't have the permit and don't can drive a ride.

Where do the other people get a ride?

Friends, family.

In some sense, it's almost unfortunate that we don't allow hitchhiking anymore.

I mean, I guess it's not allowed, certainly frowned upon.

Until I was, I don't know what.

Probably 21 or leave.

I don't know until I guess my wife, my wife had a car.

So that was my first car.

Otherwise I was hitchhiking.

Thank goodness some people picked me up.

I had a little sign.

People at least try to make big enough. So is there a comment?

They say, hey, you know, is this guy heading in my direction?

What if I would have had a phone?

Well, I guess it's because darn it, the car has been made so inexpensive, so available.

And there have been so many people out there willing to give their friends and family rides.

We didn't have really needed this.

Boy, if you don't have a friend or a family to give you a ride, you're walking.

That's pretty tough.

Now the opportunity to not go out there and, you know, Uber and Lyft are fantastic.

Except the person deserves to be paid.

Who gives you the ride?

It's not a friend.

It's not family.

You don't have a quick pro quo.

So therefore, you know, sorry to be preaching about this, but that's the situation.

And this is what this machine can unlock.

And the secret is it can unlock it for two reasons and do it affordably for those two reasons.

One is the amount of labor per trip that you give for per ride that you give ends up being really small.

If, if, if there's always ifs, if you're able to scale this thing.

And the interesting thing about this ability to scale is what you want is you want your marginal cost to be in this direction, not this direction.

And the beauty of one of the beauties of this thing is, is that it has the marginal cost in this direction.

So that the additional rider that you put onto this, it gets cheaper.

That just forces growth.

If there's a demand, if people could, if people need a ride, if people don't need a ride, doesn't do you any good.

You can't scale.

Nobody cares.

But in this situation, there is a latent demand out there of all the people that for which they haven't been permitted.

They need to bum a ride from family or friends or they can't afford.

Okay.

That is a non-trivial if there are a billion, billion person trips in the United States every day.

My estimate and I wish, I wish, you know, DOT, somebody, some research or someplace, you know, gave a more precise estimate of this.

There are at least 10% 100 million trips every day for which somebody could go some, wants to go someplace, can't get there because they can't get a ride.

And they can't afford what's out there.

At least all of a sudden, if you have a machine that can give you the ride, and in fact has this as its marginal cost.

The supply demand curve goes nuts.

It might be a good time to bring up that the first, if we want to get into this at all, but the first purposely built vehicle, the cyber cab, the first production unit has arrived.

It's going to be a while before there's real any kind of mass production, apparently, but it's a purposely built vehicle without steering wheel pedals, just seats designed to provide mobility.

Well, you know, I'll say you don't, hey, it doesn't matter if there's a steering wheel or whatever. I mean, you don't have to use it. You can uncouple it.

It doesn't need to, you know, you do need the actuators, probably two of them. In case one of them, you have another one. You need the sensors, you need the computer, and you need the software.

Now, whether or not you have a, hey, sure, we would love to have it designed to do the, the, the, the, the, the, the, the, the, we can get started much more easily. We don't have to wait for that.

We really don't have to wait for the special bill vehicle because, you know, the special bill vehicle, hey, if we go through and because of what the manufacturers do in terms of at least one manufacturer, if not more than one manufacturer does in terms of

adhering to the, to the, to the conditions that are in these regulations, they have the wherewithal to do that. Then all of a sudden, an operator can go out there and take those and do the vehicle management and all the whatever

optimization to do, to do, to increase productivity of the vehicles, deliver that respond, deliver the level of service, because it has to be delivered again.

These things aren't calling them autonomous vehicles is silly. They're not autonomous.

The whole idea is to tell them, hey, Fred needs a ride or for Fred to say, hey, I need a ride.

And for an operator to say, okay, here's your latte, you wanted a latte, you can have it.

But the latte was made because you wanted it. When you want it, the ride is given when you want it, just like, you know, if it was your own.

All of a sudden, you get to go. Yesterday, I went and got my haircut.

It made me happy. Or maybe it made my wife happy. I don't know. But of course, if it makes my wife happy, it makes me happy. So made me happy.

Just get my haircut. I had to go someplace farther than I can walk.

Well, luckily enough, I can't so drive, but I'm not driving anymore, because I'm not good at it.

I've decided at least for myself. Thank you, you're better than me, you do it.

For me, my own personal, not suggesting it to anybody, you get to make up your own mind, we're all different.

And you're right for you. Please let me be right for me. And regulations apparently let me be right for me. For me.

Thank you. These regulations are out there, and that's why I supported them.

And basically, my testimony, which is linked in my e-letter, and I guess you can look it up, they haven't posted it yet on the NHTSA site.

So I don't know. I guess I'll have to call them Monday and say maybe I didn't hit the right button and submit.

I mean, I don't know how to use computers. So it's beyond me.

But it's out there. And the basic thing is one, it's just really good what is going on now to have a serious discussion among adults about this.

To really look at the safety question as Michael suggested, looking at this as a solution to the safety problem is just, I don't know, sure.

But it doesn't have an opportunity to do that until those that are misbehaving in the current system decide to stop driving and start riding.

And I don't know what kind of regulations and so on. One's going to be able to put on these to convince. That's a personal decision.

That's a personal, private decision. It's freedom of an individual to choose.

That's America. Freedom. Chances of nits is telling any of us, you do this.

Not everybody puts on a seatbelt now.

That's why I thank goodness for airbags that are passing. You don't have to turn them on.

Be nice if the seatbelt, I mean, they've tried to make them automatic and I mean, it'd be great if it could.

Then some people are probably going to cut them anyway or figure it out or this on the web someplace will be a way to the game the system.

But that's that's what freedoms are all about. Individual choice.

And, you know, hey, I think you should all be everybody should be riding, of course.

But the fact that now the availability of this the availability of an operator to go out and provide the mobility now, the folks who now be able to go get their hair cut.

Go visit a friend.

Go to the hospital. Go to the doctor's appointment. Go to go to work.

Be able to work while you go to work.

I mean, that's, you know, all of us academics have loved. Oh, my goodness, I'll be able to work while I go to work.

I don't know.

I gain. No, I'm not a gamer. So of course I would work.

But the enjoy driving is really work.

It's amazing when you're riding shotgun how when somebody else is driving, how you can observe the world and just feel better about the world, but what's out there instead of focusing on the road ahead.

With your, with your, you know, oh, your display system showing you how fast you're going, you know, in big letters right there in your field division, you don't even have to look down the speedometer.

I mean, just to me, really, I wouldn't look at my speed all the time.

This is why I always love the even stupid cruise control. Boy, I mean, you know, okay, I want to go this speed, put it on there. I mean, for me to sit there, am I, am I right on it?

Hey, the interest as a side note, the interesting thing about AI.

It's forgetting.

One of the things the brain does really well.

Forget.

Can you imagine if we didn't forget?

We go nuts.

How many sensors do we have all over our bodies?

Beauty of this thing. It blows up much of that information away.

It focuses on the beautiful things.

Really ugly things.

And, you know, for, certainly, I'm sure everybody, you know,

you really, the way you grow a tough skin,

as you forget about some of the ugly stuff,

forgetting is really important.

But anyway, I digress.

You have some links in the latest newsletter to a few AI pieces as well.

You know, well, because of course it's relevant.

All we talk about because, because as Michael pointed out, while driving is easy.

Well, the way I like to put it is if everybody behaved,

driving is almost trivial.

The only, I don't know, whatever, 90% some overwhelming percent of the challenges

are associated with driver misbehavior.

It is really pretty easy for us to design a system that stays right down the center of a lane.

Especially if the road surface is smooth and there's paint, trivial.

But what do you do when whatever the myriad of things that are occurring in that, in that lane,

that shouldn't be occurring in that lane, shouldn't be.

But they is, whoa.

And so that ends up to be in the real challenge.

And then, you know, if you say, well, it's got to do all those things.

Well, then, you know, that's when I give up.

I know can't do all can do many of the deal with many of the misbehaviors that are out there.

And that's where get to the crash mitigation piece of this darn thing.

Yeah. How do you deal? How do you, how does society deal with the misbehaviors?

Deals with crash mitigation, putting, putting basically, you know, pillows around.

My goodness, to me, one of the fundamental values of all this, this new stuff.

It just does it better than I do.

Now, we have to be careful because it probably doesn't do some things as well as I do.

I haven't been able to identify one yet, but maybe it does.

I guess I took driver Ed so long ago that I've forgotten what to do if there is a car heading towards me in my lane on a two lane road.

Oh, I really don't. I don't think I cut that class. I don't think they ever taught that.

I don't think they teach it now.

I don't know what to do.

I've always refused to drive in England and other places where they drive on the wrong side of the road.

Why? Because if I'm confronted with something,

my knee jerk reaction is going to be related to my experiences, which is driving on the right side of the road.

If I'm confronted with this, I'm supposed to be on the wrong side of the road.

What are the chances I'm going to do the right thing?

I think I'm in trouble here.

The house is going to win on that bet. I have no chance.

I mean, the spread between the house and me is so large.

They're not rewarding me with an appropriate pay. I'm a dead loser, probably dead and a loser.

You don't need to do arithmetic to do that one.

And so in these situations, I don't know, I'm thinking the machine because it can react faster, because it can see better, because it saw it earlier, because of the opportunities it has there.

I think I'm the house then. I'm not the chunk better.

But that's, hey, that's my opinion. So I don't know how do we get that? I mean, supposedly these AI things hallucinate.

There are situations I think when I've been driving that I did one of these guys, because I hallucinated.

I didn't somehow like whatever reflection to do the duh who knows what combination duh duh duh duh. It's kind of lost.

Luckily, so far I've been able to regain.

I had, I don't think there's any evidence of these things having sustained hallucinations.

Where because because everything is happening dynamics, everything, you know, it's looking at instant data.

But then the next 30th of a second, those data are different.

And the next 30th of a second, they're different because the future is always different.

I don't know. Aristotle almost said that or something.

It is different.

And so now all the stuff going in there, if there is a lip on one, I don't think it's been demonstrated there is sustained hallucinations.

Sustained over what? I don't know.

10th of a second.

Even then, can you recover from that?

Essentially, sufficiently number of times.

I don't think anybody's demonstrated that they don't.

So everyone wants them up.

I get stunned.

Luckily, I guess because I'm still here, I recovered.

So in the context of, you know, extreme detail, mathematical and analytical stuff.

I don't know. I think we can be the house on this one instead of the charm.

Interesting way to put it.

Well, I mean, that's the way I think all my students think they can develop algorithms to be betting. I don't know, because apparently, you know, a couple of guys who knows when and MIT did it.

I don't know if they cheated or not, whatever. I'll assume they didn't. But okay, one occurrence.

Before we wrap up, Alan, did you want to comment on the this week in space flight piece that you?

Yeah. Yeah. Just a little bit. Yeah.

Yeah. Well, you know, I think Melissa Segal, again, that's something that I watch every week in terms of what's happening in space flight. But again,

This is on NASA's Starlight.

Yeah. One has to, a lot of things are very hard. And we don't necessarily perform quite as well as we should.

But I think we should also have the opportunity to do the improvements. And as we learn, we fix. We don't know what we don't know.

That's tough. But we should have on top of that as soon as we find out.

Let's fix it.

Because then we don't have the excuse of we didn't know. And we didn't know is not just the dog ate my homework.

Okay, it is, it is a fundamental frailty of the human species, which I guess I'd like to say is the beauty of the human species.

Because of course, if the future, if you could forecast the future, but wouldn't be worth living it.

Just again, the way I look at things, you look at things however you want, please, I encourage you, I don't want you to look at things necessarily the way I look at them.

But that one, you know, what's exciting. I mean, has has US played Canada yet for the golden in hockey?

The women one yesterday. Is that tomorrow? Oh, yeah. Oh, my goodness. I can't wait. Oh, my former Canadian hockey student friends.

If we if we win that one too.

Oh, they're going to hear it from me.

Yeah.

Thanks for spending time with us and stay safe.

Smart Driving Cars episode 407- New Automated Driving Regs

collision mitigation

anti-lock brakes

seat belts

airbags

crush zones

autonomous cars

WP29

Tesla Cybercab

About this episode

Original notes