The Coming Disruption of Transport

[afransen]

The thing that I struggle to understand is why AEVs wouldn't lead to more intermodal shipping. If you remove the driver and lower fuel costs, then return is entirely about asset utilization. Why send one container on one truck across the country when you can have that truck pick up containers and drop them off at the transloading terminal in town all day, increasing utilization?

There may be high value JIT goods that justify a single truck driving across the continent by itself. But for a lot of commodities, I imagine rail and trucking companies can leverage automation and lower fuel cost to improve the speed and efficiency of transloaded shipments across North America.

The closer we get to full automation, the closer we get to physics dictating operating costs. And there will be some threshold distance where steel on steel beats rubber on asphalt.

Intermodal is a function of the spread between the cost of shipping by rail vs road. I think it's probably fair to say that rail will always be slower unless really dramatically change (cross-continental high speed freight rail?).

CN clearly thinks that autonomous trucks are a relevant business to them, perhaps as a complement or competitive threat. They are invested in TuSimple.

 

[afransen]

In other news, everyone's favourite gadgetbahn/egotistical billionaire vanity project is moving ahead in Las Vegas. The Las Vegas City Council has given approval to expand The Boring Company's Vegas Loop system to 55 stations and 35 miles, connecting the existing system to the airport, Allegiant Stadium and downtown LV. The system will be privately financed by TBC and destinations along the route. TBH hopes to begin excavation next year.

Sources:

Elon Musk's Boring Company will build a 34-mile tunnel network underneath Las Vegas

The company receives approval to expand its tunnels underneath the city.

www.engadget.com

Elon’s One-Lane Vegas Tesla Tunnels Expanding Into Full Boondoggle

The dodgy Las Vegas tunnel network will now be over 34 miles long with at least 55 stations.

gizmodo.com

 

[TRONto]

In other news, everyone's favourite gadgetbahn/egotistical billionaire vanity project is moving ahead in Las Vegas. The Las Vegas City Council has given approval to expand The Boring Company's Vegas Loop system to 55 stations and 35 miles, connecting the existing system to the airport, Allegiant Stadium and downtown LV. The system will be privately financed by TBC and destinations along the route. TBH hopes to begin excavation next year.

Sources:

Elon Musk's Boring Company will build a 34-mile tunnel network underneath Las Vegas

The company receives approval to expand its tunnels underneath the city.

www.engadget.com

Elon’s One-Lane Vegas Tesla Tunnels Expanding Into Full Boondoggle

The dodgy Las Vegas tunnel network will now be over 34 miles long with at least 55 stations.

gizmodo.com

Hopefully LV will be able to use the tunnels for shuttle buses if/when the TBC ditches the project.

 

[afransen]

Hopefully LV will be able to use the tunnels for shuttle buses if/when the TBC ditches the project.

As long as those shuttle buses are the size of cars!

 

[kEiThZ]

Intermodal is a function of the spread between the cost of shipping by rail vs road. I think it's probably fair to say that rail will always be slower unless really dramatically change (cross-continental high speed freight rail?).

CN clearly thinks that autonomous trucks are a relevant business to them, perhaps as a complement or competitive threat. They are invested in TuSimple.

Rail will always be slower. Sure. But it will also be cheaper. Labour and fuel aren't the only costs involved in transport. And not everything requires rapid and urgent delivery. Indeed, time in transit is often used as virtual warehousing by industry for items that aren't very time sensitive.

I do think CN interested in TuSimple exactly because of my logic. Enabling cheaper last mile delivery, only helps their business.

 

[C_Johnson_1995]

In other news, everyone's favourite gadgetbahn/egotistical billionaire vanity project is moving ahead in Las Vegas. The Las Vegas City Council has given approval to expand The Boring Company's Vegas Loop system to 55 stations and 35 miles, connecting the existing system to the airport, Allegiant Stadium and downtown LV. The system will be privately financed by TBC and destinations along the route. TBH hopes to begin excavation next year.

Sources:

Elon Musk's Boring Company will build a 34-mile tunnel network underneath Las Vegas

The company receives approval to expand its tunnels underneath the city.

www.engadget.com

Elon’s One-Lane Vegas Tesla Tunnels Expanding Into Full Boondoggle

The dodgy Las Vegas tunnel network will now be over 34 miles long with at least 55 stations.

gizmodo.com

I feel like people use boondoggle vanity projects like these to excuse their dismissal of autonomous vehicle deployment in general. Tesla and AVs are a false equivilance and the sooner Musk is deposed from the CEO position, the better.

In other AV news:

MSN

www.msn.com

I really hope Toronto is on the list of expansion cities. Its streets, especially in the outlying areas would be very conducive to such a service and the test in a colder climate would be useful in building resiliencies into the system. Fingers crossed.

Rail will always be slower. Sure. But it will also be cheaper. Labour and fuel aren't the only costs involved in transport. And not everything requires rapid and urgent delivery. Indeed, time in transit is often used as virtual warehousing by industry for items that aren't very time sensitive.

I do think CN interested in TuSimple exactly because of my logic. Enabling cheaper last mile delivery, only helps their business.

The problem is that rail has very limited avenues to automate operations due to long stopping distances, limited sightlines and unpredictable nature of line obstructions. Therefore, while it is likely that trucking costs will drop by half, rail will no such reduction.

The important part also is that being cheaper isn't the important part. It is being the option with the most value. Trucks handle the majority of freight in Canada and the United States and this is because shippers are willing to sacrefice up front cost for reliability, individualized service and speed. These three features of trucking will, when considered as a whole, often make the total cost in terms of overall value cheaper. And even more than that, autonomous trucks will make these three factors even better.

So what we have coming are trucks which can move further faster due to no rest requirements, and will likely be more reliable, but also the cost differential will be massively cut. For probably about half of rail shippers, switching to truck will be a no-brainer.

The outcome of autonomous truck adoption will likely be nearly identical to that of the implementation of the national highway networks. This includes the shifting of industrial geographies and practices which bake trucking into the supply chain with no easy substitute. The difference is that railways are at a much more depressed state of importance than they were when trucking was taking off and dont have a lot of slack in terms of profitability and demand.

The solution is clear. Nationalize. Manage the decline. Stem the bleeding. Attempt a partial recovery.

I think it is fair to say that we wont likely ever see railways as prominent or powerful as we see them today again, but it is still possible to keep them as part of the transport mix, but we need radical change.

 

[lenaitch]

I feel like people use boondoggle vanity projects like these to excuse their dismissal of autonomous vehicle deployment in general. Tesla and AVs are a false equivilance and the sooner Musk is deposed from the CEO position, the better.

In other AV news:

MSN

www.msn.com

I really hope Toronto is on the list of expansion cities. Its streets, especially in the outlying areas would be very conducive to such a service and the test in a colder climate would be useful in building resiliencies into the system. Fingers crossed.


The problem is that rail has very limited avenues to automate operations due to long stopping distances, limited sightlines and unpredictable nature of line obstructions. Therefore, while it is likely that trucking costs will drop by half, rail will no such reduction.

The important part also is that being cheaper isn't the important part. It is being the option with the most value. Trucks handle the majority of freight in Canada and the United States and this is because shippers are willing to sacrefice up front cost for reliability, individualized service and speed. These three features of trucking will, when considered as a whole, often make the total cost in terms of overall value cheaper. And even more than that, autonomous trucks will make these three factors even better.

So what we have coming are trucks which can move further faster due to no rest requirements, and will likely be more reliable, but also the cost differential will be massively cut. For probably about half of rail shippers, switching to truck will be a no-brainer.

The outcome of autonomous truck adoption will likely be nearly identical to that of the implementation of the national highway networks. This includes the shifting of industrial geographies and practices which bake trucking into the supply chain with no easy substitute. The difference is that railways are at a much more depressed state of importance than they were when trucking was taking off and dont have a lot of slack in terms of profitability and demand.

The solution is clear. Nationalize. Manage the decline. Stem the bleeding. Attempt a partial recovery.

I think it is fair to say that we wont likely ever see railways as prominent or powerful as we see them today again, but it is still possible to keep them as part of the transport mix, but we need radical change.

Anticipation may have to be tempered. From the article:

"eventually enter new cities across the state"

 

[crs1026]

The problem is that rail has very limited avenues to automate operations due to long stopping distances, limited sightlines and unpredictable nature of line obstructions.

How do you explain that trains operate just fine today without automation - with the human drivers having exactly the same constraints and obstacles?

A railway line is not a very dynamic environment. It is very easy to detect new or changed items that may require a response from the train, because the physical environment is mostly fixed in place. Even road crossings, which represent the most frequently changing stimulus, and the greatest source of risk, can be equipped with sensors - as is already the case on some Amtrak lines.

Therefore, while it is likely that trucking costs will drop by half, rail will no such reduction.

That math only works if labour represents 50% of trucking costs, and if fuel and capital costs for trucks don’t rise. And if railways do automate or find efficiencies?

A large increase in truck traffic will necessitate new investment in roads. That cost will be passed to the industry. And create huge public controversies. The solutions will not be cheap. I wonder, also, when AV developers will recover the investment to date. Amortising development costs will represent a new cost for trucking licensees. The technology is not being developed for free.

The important part also is that being cheaper isn't the important part. It is being the option with the most value. Trucks handle the majority of freight in Canada and the United States and this is because shippers are willing to sacrefice up front cost for reliability, individualized service and speed. These three features of trucking will, when considered as a whole, often make the total cost in terms of overall value cheaper. And even more than that, autonomous trucks will make these three factors even better.

I agree that railways need to pull up their socks here.

IMHO the railways are bobbling the automation file, as demonstrated by the recent hearings on one person crews in the US. But that’s the value of competition… if, as some suggest, trucking costs do fall, railways will face pressure to change their approaches.

But - if railways can achieve reliability, how much freight actually needs the added speed that highway offers? If a component or raw material is needed immediately, then it will go by truck. But the cost of many goods being in transit for an extra day may not be that compelling so long as the promised delivery tcd is met.

The outcome of autonomous truck adoption will likely be nearly identical to that of the implementation of the national highway networks. This includes the shifting of industrial geographies and practices which bake trucking into the supply chain with no easy substitute.

There will not be a National Highway Network 2.0 created at the public’s expense.

What will be interesting is to watch truckers and railroaders battle out the legislative and regulatory decisions that will result from AV introduction. Both are awfully good at lobbying.

- Paul

 

[kEiThZ]

There's a lot of folks here who are tech fans, who seem to have never talked to anybody actually working in logistics. Time to ship is not the only consideration.

Also, I think a lot of people are going to be surprised by shipping costs even without the driver. To start with, tires don't get cheaper. Especially if the roads get congested and filled with potholes.

 

[C_Johnson_1995]

How do you explain that trains operate just fine today without automation - with the human drivers having exactly the same constraints and obstacles?

A railway line is not a very dynamic environment. It is very easy to detect new or changed items that may require a response from the train, because the physical environment is mostly fixed in place. Even road crossings, which represent the most frequently changing stimulus, and the greatest source of risk, can be equipped with sensors - as is already the case on some Amtrak lines.

It gives liability in the event of an accident to an individual. The problem is creating perception systems for automating mainline railways which dont give too many false stoppages. Autonomous road transport is much less demanding on perception and has a nearly two decade head start.

There will not be a National Highway Network 2.0 created at the public’s expense.

What will be interesting is to watch truckers and railroaders battle out the legislative and regulatory decisions that will result from AV introduction. Both are awfully good at lobbying.

That's not what I'm arguing. What I am saying is that the disruptive nature of autonomous driving will have a nearly identical affect on railways as the proliferation of autonomobiles and trucking did in the 1950s-1970s.

As for lobbying, I think trucking will likely have the upper hand as they continue to do today. Generally, most railway lobbying has been to the effect of seeking permission to abandon services.

There's a lot of folks here who are tech fans, who seem to have never talked to anybody actually working in logistics. Time to ship is not the only consideration.

I never argued transit time was THE dominant factor, but trucks are generally able to offer a level of service and reliability that railways will likely never be able to match, and that lead will only widen with autonomy. Therefore, when trucks become cheaper, they will become more attractive to a larger subset of rail shippers looking for reliability and service.

The other important thing to remember is that most modern industrial developmemts are not along railway lines and are truck dependent, so a large portion of shippers dont have rail as an option.

For the record, while I am interested in the tech, I am horriffied at its potential to create disruption and send us in the wrong direction which is why I am keen to stimulate discourse to better allow it to coexist in the modern transport ecosystem.

The one thing I have been thinking about for Canada is if we should go toward something of a UK freight model whereupon traditional carload services are discontinued in favour of block trains and intermodal. This would be able to keep the network moving more fluidly and leverage the inherent efficiencies of rail and avoid time and labour intensive customer delivery and marshalling. (Obviously this is something of a spitball as it is easier said than done, though it is interesting to consider the merits)

 

[kEiThZ]

Therefore, when trucks become cheaper, they will become more attractive to a larger subset of rail shippers looking for reliability and service.

Trucking becoming cheaper is based on two cost reductions:

1) Labour -Elimination of the driver
2) Fuel - Electrification reduces the cost of fossil fuels.

But this ignores the capital cost and maintenance costs of the vehicle itself. And they don't go away. In some cases, they actually increase. Just look up how much tires cost and how long they last.

Also, "reliability and service" is really not a consideration for most things shipped. Somebody shipping perishable goods or high value goods might care. But somebody shipping a container of wheat isn't likely to care whether it is shipped in 24 or 48 hrs.. And a lot of the railways' business is commodities. They most definitely will not pay more to get marginal increases in reliability.

Also, there's still zero proof offered that railways will be categorically excluded from using the same technology to make improvements and reduces their costs further.

 

[crs1026]

It gives liability in the event of an accident to an individual. The problem is creating perception systems for automating mainline railways which dont give too many false stoppages. Autonomous road transport is much less demanding on perception and has a nearly two decade head start.

I would argue the exact opposite. Sensing systems can already differentiate between a moose, a human, and a dog. And spot and assess any physical object that deviates from the known physical layout - eg a shopping cart or an auto abandoned on the tracks. Detection of track conditions, switch positions, etc is already automated. The only key constraints are telecom bandwidth to permit remote oversight, and physical access/response time to malfunctions out along the line. These things are solvable and affordable, if railway executives would look a bit further than this quarter’s operating ratio.

A truck has to manage a dynamic environmeant of other vehicles, pedestrians, road alterations, emergency vehicles…. and differences in road surface as built, plus variability created by construction, detours, weather, etc. Automated trains know exactly where the tracks are, and how much to slow down on curves. The AI required to navigate on a railway is much less.

The other important thing to remember is that most modern industrial developmemts are not along railway lines and are truck dependent, so a large portion of shippers dont have rail as an option.

As noted earlier, drayage is exactly where AV’s will meet the railways. With automation, a greater number of hubs and transfer points can be created.

The one thing I have been thinking about for Canada is if we should go toward something of a UK freight model whereupon traditional carload services are discontinued in favour of block trains and intermodal. This would be able to keep the network moving more fluidly and leverage the inherent efficiencies of rail and avoid time and labour intensive customer delivery and marshalling. (Obviously this is something of a spitball as it is easier said than done, though it is interesting to consider the merits)

My theory is that automation of railways will enable a return to shorter more frequent trains and more flexible operating patterns that create opportunity to serve more customers more flexibly.

Already, DPU trains are effectively two or three shorter trains lashed together. The tradeoff is lower labour costs, but at the expense of a very inflexible operation - unloading a 14,000 foot container train, let alone marshalling it and testing the brakes, is what creates the low velocity. The railways have shed local service because lifting and setting off at midpoints using trains of that length is cumbersome and inefficient.

If we reach a point where trains again fit the existing sidings, instead of having to plan meets and overtakings from a hundred or more miles away (due to the dearth of long sidings), we can undo that disruption.

For bulk commodities, travel time is already constrained by maritime terminal capacity, transloading time and delays at sea, etc. The low velocity creates cost via added fleet size and need for on-route capacity, yes, but rushing grain to port when the ship can’t load for another week does not add value.

- Paul

 

[afransen]

Trucking becoming cheaper is based on two cost reductions:

1) Labour -Elimination of the driver
2) Fuel - Electrification reduces the cost of fossil fuels.

But this ignores the capital cost and maintenance costs of the vehicle itself. And they don't go away. In some cases, they actually increase. Just look up how much tires cost and how long they last.

Also, "reliability and service" is really not a consideration for most things shipped. Somebody shipping perishable goods or high value goods might care. But somebody shipping a container of wheat isn't likely to care whether it is shipped in 24 or 48 hrs.. And a lot of the railways' business is commodities. They most definitely will not pay more to get marginal increases in reliability.

Also, there's still zero proof offered that railways will be categorically excluded from using the same technology to make improvements and reduces their costs further.

Labour and fuel are probably about 2/3 the running cost of a truck. Saying that the cost can 'only' drop by two thirds seems to be missing the point. Rail can benefit from some of the same technologies, but I don't think the scope for cost reduction is as large since rail is already more efficient.

I agree that commodities are going to continue to be shipped by rail. They are fungible and have relatively low embodied value than manufactured goods that are typically shipped intermodal.

 

[kEiThZ]

I agree that commodities are going to continue to be shipped by rail. They are fungible and have relatively low embodied value than manufactured goods that are typically shipped intermodal.

And if commodities continue to be shipped by rail, this whole fantasy of the railways dying falls apart.

People who think autonomous trucks will kill rail have the false idea what all cargo is equally important and therefore will shift to a faster and more responsive transport system. However, most cargo is just not that time sensitive. And most of what is, already travels by truck. The exact rail-truck share split might change a little. But it's not going to result in the railways being bankrupted and needing nationalization as some of the crazy speculation on here suggests.

 

[afransen]

I agree that that is likely wrong. There are Alan Fisher rail fan types who think it's feasible to go back to box cars delivering Barbies and tomatoes into cities who are just as wrong in the other direction. I wonder how much losing intermodal freight volumes will hurt the financial performance of railways, but I don't think they are going out of business altogether. If anything, falling coal demand is a bigger threat. If we are remotely serious about climate goals, coal is going to near zero. Even if we aren't, renewables and natural gas will squeeze out a lot of coal demand.