Hyperlanes, Bullet Trains, And AI Self-Driving Cars

Dr. Lance Eliot, AI Insider

Image for post
Image for post

[Ed. Note: For reader’s interested in Dr. Eliot’s ongoing business analyses about the advent of self-driving cars, see his online Forbes column: https://forbes.com/sites/lanceeliot/]

I feel the need, the need for Maglev speed.

The Maglev has been considered the fastest commercial High-Speed Rail (HSR) line and whisks passengers at a breathtaking 267 miles per hour from the Pudong airport to the Longyang station in Shanghai, a distance just shy of 20 miles.

Named the Maglev because it uses magnetic levitation, it has been a marvel since it first opened in 2004.

Let’s call high-speed rail lines a more flavorful name, bullet trains.

If you hold your breath, you might get a chance to someday ride a bullet train in California.

That’s actually a funny statement because anyone that lives in California knows that we’ve been pining away to have a bullet train for quite a long time. You’d need to have a large pair of lungs to hold your breath for as long as we’ve contemplated having our very own California bullet train.

In 2008, California residents voted in favor of Proposition 1A, a $9 billion bond to help kick-start a bullet train that would run from Los Angeles to San Francisco. Nearly seven years later, in 2015, there was an initial ground breaking ceremony that took place in Fresno, California.

Some liken the bullet train to waiting for Godot.

It might or might not ever really happen.

If it does happen, the guess is that it will be sometime in the 2030s before it is fully operational.

The distance involved is quite a bit further than the Maglev that I mentioned earlier — the California bullet train will go a distance of about 438 miles.

Recall that the Maglev went a scant 20 miles or so.

When you consider a bullet train and what it is, you need to be aware of the distance it goes and the intended average speed.

The terrain makes a difference too (such as, for California, the envisioned bullet train will need to somehow deal with mountains, rivers, lakes, etc.).

And believe it or not, the bullet train in California is supposed to route through five of the ten most major cities in the entire state.

Rights of way become an issue.

Protecting the track from people is an issue.

Protecting the track from animals is an issue.

The headaches abound.

One ongoing question involves how the bullet train in California will impact the travelers going between the two primary endpoints or locations in terms of mode of travel.

You need to know that a flight from LAX to San Francisco is usually about an hour and a half in duration, which is important when considering the bullet train alternative.

Business travelers that are not especially price sensitive are going to focus not so much on the lower cost presumably of the bullet train travel, but how long it takes them to make the distance.

To-date, the California bullet train authority has said that it will be a time of 2 hours and 40 minutes. Thus, it would apparently take about an hour longer via the bullet train than flying.

Bullet train advocates say that besides reducing the number of flights, the advent of the bullet train will pull people out of their polluting gas-guzzling cars too.

They claim that besides the environmental improvement, the cost per passenger per mile is going to be much better for the bullet train than going via conventional car.

Car Versus Bullet Train

The maximum speed limit for cars in California is legally stated as 65 mph, though in some areas it is allowed for drivers to go 70 mph.

If you take the 438-mile distance of the proposed bullet train, and divide it by 70 mph, the projected time by driving a car would be 6 hours.

You could argue that you are unlikely to be able to go 70 mph that entire distance. Perhaps due to traffic. Perhaps due to debris on the roadway. Perhaps due to foul weather. And so on.

Let’s just assume for the moment that you could go 70 mph the whole way (which, actually is more often possible than one might assume), taking about six to seven hours or so.

One advantage of taking your car is that you are able to go from door to door.

Suppose you could somehow maintain an operating speed of 85 mph on the 438 miles distance.

This would mean that you’d get to San Francisco from LA in about 5 hours.

Suppose you were willing to edge your car up to 100 mph?

The distance could be covered in approximately 4 ½ hours.

Plus, that’s with door to door convenience.

AI Self-Driving Cars Versus Bullet Train

What does this have to do with AI self-driving driverless autonomous cars?

At the Cybernetic AI Self-Driving Car Institute, we are developing AI software for self-driving cars. One aspect of the future for self-driving cars will be their ability to operate in a high-speed long-distance mode.

Let’s consider what you could do in terms of using an AI self-driving car to make the ride that the California bullet train is proposing to do.

First, assuming that the AI self-driving car is a true Level 5 self-driving car, you don’t need to worry about the driving aspects. The point being that with a true AI self-driving car, you would not need to be driving the LA/SF and could enjoy the ride as a passenger, similar to what you would do on the bullet train or on an airplane.

In that sense, the AI self-driving car is certainly superior to a conventional car that requires a human driver and so the comparison of the bullet train to a conventional car differs with an AI self-driving car.

On a related aspect, it is anticipated that most AI self-driving cars will be Electrical Vehicles (EVs). This does not need to be the case, but it’s highly likely. The reason I mention this aspect is that the bullet train advocates emphasize that gas guzzling cars are polluters.

An AI self-driving car that is based on an EV platform will take away that argument and thus remove the pollution differences consideration.

Someone that is reading this and perhaps is a bullet train advocate will probably say that I’ve just tried to pull a fast one about those EVs. Most EVs today average maybe 200 miles distance before a recharge is needed. Thus, trying to drive the 438 miles of LA/SF would require about 4 stops to do a series of needed recharges.

Even a fast recharger will still add a lot of time to the total time of having the AI self-driving car get you the distance involved after all of those needed stops (a faster alternative is a battery swap, though this is only available on a limited basis currently).

I grant you that today’s EVs have a distance capability that is more suited to short-distance hops, such as the average daily commute to work and home.

But, please note that the Chevy Bolt can get about 238 miles on a single charge. The Tesla Model S can get about 337 miles.

Those seem like outliers now, but the EV makers all realize that without being able to cover larger distances, the masses aren’t going to be willing to buy EVs.

Remember that the bullet train is not expected to be in use until the 2030s.

Give the EV makers a few years from now and we’ll likely see vast improvements in distance coverage, long before the 2030s.

How Fast Are We Willing To Have Cars Go

The biggest catch might be the speed aspects.

Do we want these true AI self-driving cars to be going at high speeds?

What is the boundary we are comfortable with? Will maintaining a speed of 85 mph be acceptable, which seems like it would since it’s pretty much the norm now anyway. Some would say they see nothing wrong with aiming at a speed of 100 mph. Others would push the envelope and argue that 125 mph would be acceptable.

At a speed of 125 mph, the 438 mile distance of SF/LA would take about 3 ½ hours.

That’s getting pretty darned close to the supposed 2 hours and 40 minutes for the bullet train.

As an aside, there are many that argue with the alleged 2 hours and 40 minutes by the bullet train authority, and point out that this is only based on computer based simulations of what might be possible.

And, it is suggested that those simulations have questionable assumptions about what the bullet train will really do when encountering the various aspects of how the bullet train is really going to be constructed.

You might also find of interest the price tag for the California bullet train.

The price has been going up each passing year, apparently due to re-estimating and also due to deeper explorations about what the construction will consist of. The price tag had been around $33 billion. Now, it’s ballooned to about $77 billion. Some say there is no end in sight.

For AI self-driving cars, you might say that there is no construction cost per se, since these AI self-driving cars are going to be able to traverse the distance without any kind of special capability.

In essence, an everyday true AI self-driving car can make the drive.

We could though try to level the playing field by looking at the chances of accidents and the safety factor.

Bullet trains have been remarkably safe.

Very few accidents.

Would the true AI self-driving cars be able to match that kind of safety record?

All in all, it would seem that the I-5 might have some troubles trying to ensure that true AI self-driving cars could go along at an operating speed of say 100 mph or higher, not due to the road itself and not due to debris, but more so to the traffic control.

That being said, one of the potential aspects about AI self-driving cars involves their ability to communicate with each other using V2V (vehicle to vehicle communications), and possibly V2I (vehicle to infrastructure communications).

As such, the AI self-driving cars could communicate electronically to coordinate their movement on the I-5. This would presumably allow for speedy lane changes and avoid lane blockages.

In terms of the chances of an AI self-driving car getting involved in a domino-like car accident, this is something that still could happen, even with true AI self-driving cars using V2V and V2I.

One important additional consideration is the mixture of human drivers and AI self-driving cars on the roadways.

As I’ve mentioned many times, we are not going to anytime soon have only AI self-driving cars on our roads.

As such, we are then in a bit of a predicament or conundrum because it seems as though true AI self-driving cars could potentially be a viable alternative to or complimentary to a bullet train, but not if we can’t do something about the traffic aspects.

You could opt to dictate that only true AI self-driving cars could go on the I-5.

This is likely to create a widespread public backlash. It would be perceived that those elitist-owned AI self-driving cars have taken over a key public roadway and a vital connector between northern and southern California. This might be hard to convince the public at large to accept.

Use Of Hyperlanes

Another approach would be to “transform” the I-5 into a so-called hyperlane.

This might consist of expanding the number of lanes and dedicating some lanes for exclusive use by the true AI self-driving cars. The remaining lanes could be used by conventional car traffic. This dual lane division would separate the two, allowing the speedy and V2V communicating AI self-driving cars to use their own space for trying to maintain top operating speeds. Conventional car drivers would have less to complain about since they presumably would still be able to drive on the I-5.

The lane expansion might be to widen the existing road, as mentioned above. Or, some might argue to go over the top of the existing lanes and provide an aerial or raised alternative. Going underground beneath the existing road would seem prohibitive. In whatever way it might be undertaken, there’s certainly a hefty cost to this infrastructure change. Would we better off spending the $77 billion on the bullet train or some (or all of it) on transforming the I-5 to become a hyperlane?

Should we consider having both the bullet train and the hyperlane?

This would require a belief that the traffic will be so voluminous that both approaches are needed.

Some would argue that neither approach is warranted because we’ll “soon” hopefully have flying cars or drones that can carry people.

If we can transport people in the air, doing so via individual or small group transport, would that be “better” than conventional airplanes, and/or better than a bullet train or ground-based self-driving cars? Lots of potential options, no clear cut answers.


Overall, will the advent of AI self-driving cars spell the death knell for bullet trains?

If we have zillions of true AI self-driving cars on our roadways, and if they are the energizer of the ridesharing economy, it could be that the notion of and adoption of bullet trains is no longer prudent. It’s hard to say right now whether and when we’ll have widespread true AI self-driving cars. In the meantime, the bullet train still holds promise, but only in circumstances that showcase how the bullet train is justifiable, given its myriad of infrastructure and societal requirements.

I’d advise taking a ride on a bullet train as soon as you can, in case they eventually become extinct.

For free podcast of this story, visit: http://ai-selfdriving-cars.libsyn.com/website

The podcasts are also available on Spotify, iTunes, iHeartRadio, etc.

More info about AI self-driving cars, see: www.ai-selfdriving-cars.guru

To follow Lance Eliot on Twitter: https://twitter.com/@LanceEliot

For his Forbes.com blog, see: https://forbes.com/sites/lanceeliot/

For his AI Trends blog, see: www.aitrends.com/ai-insider/

For his Medium blog, see: https://medium.com/@lance.eliot

For Dr. Eliot’s books, see: https://www.amazon.com/author/lanceeliot

Copyright © 2020 Dr. Lance B. Eliot

Written by

Dr. Lance B. Eliot is a renowned global expert on AI, Stanford Fellow at Stanford University, was a professor at USC, headed an AI Lab, top exec at a major VC.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store