The Coming Disruption of Transport

[crs1026]

I feel like New Flyer is doing it wrong. The right form factor for an AV is much smaller than a city bus because it allows for demand-responsive door to door service. In essence, that is what an AV is meant to do; replace transit like busses and light rail by shifting people out of large vehicles and concentrated corridors and putting them on many, much smaller vehicles travelling non-fixed routes.

This approach doesn’t preclude that, but it offers transit operators a familiar package that conforms to existing paradigms, and delivers a major benefit (labour cost reduction). It’s a big enough shift just to get comfortable with this change without having to adjust the entire service delivery model. That can happen later.

I don’t expect many immediate sales, but as a demonstration for both operators and customers, it can be inserted into an existing service without disrupting everything around it. The very fact that not much changes (beyond the driving) is the whole point. Seems like a good way to win over transit folks to me.

- Paul

 

[2transpo]

The very fact that not much changes (beyond the driving) is the whole point. Seems like a good way to win over transit folks to me.

I think that is the problem. Transit agencies and "urbanists" are too stuck in their ways. Transport is changing. Much in the same way that airlines are moving away from large aircraft and hubs to smaller point to point flight, so will AVs bring to land transport. It is just more efficient and provides better customer service.

Really, alot of the light rail projects we see under construction are entirely unnecessary in the era of autonomous vehicles.

The same goes for commuter/regional rail and intercity rail. That is why I am against a project like HFR. If I am able to ride in an autonomous vehicle from Toronto to Ottawa, I will never ride the train again because it will have no advantages. I will be able to rest on the journey, get work done and probably arrive directly at my destination. It would also probably be cheaper because AV providers would have very few fixed costs. I could also see it being better for the environment as these AVs would overwhelming be electric whereas VIA will still be running diesel Siemens Chargers (that is if they weren't already shut down).

 

[north-of-anything]

It's easier to maintain one long vehicle than twenty different short vehicles. Buses and electric rail are not going anywhere...

 

[MrGoose]

I think that is the problem. Transit agencies and "urbanists" are too stuck in their ways. Transport is changing. Much in the same way that airlines are moving away from large aircraft and hubs to smaller point to point flight, so will AVs bring to land transport. It is just more efficient and provides better customer service.

Really, alot of the light rail projects we see under construction are entirely unnecessary in the era of autonomous vehicles.

The same goes for commuter/regional rail and intercity rail. That is why I am against a project like HFR. If I am able to ride in an autonomous vehicle from Toronto to Ottawa, I will never ride the train again because it will have no advantages. I will be able to rest on the journey, get work done and probably arrive directly at my destination. It would also probably be cheaper because AV providers would have very few fixed costs. I could also see it being better for the environment as these AVs would overwhelming be electric whereas VIA will still be running diesel Siemens Chargers (that is if they weren't already shut down).

So do you expect everyone to take an AV from Keswick to Downtown Toronto? Sure the individual customer experience is better, but the negative externalities of a society completely reliant on autonomous vehicles are too great to bear. Urban sprawl for hundreds of kilometers in every direction. A blasted wasteland of parking garages and highways decimating our city centers. 20 Lane highways ploughing their way across all of Toronto.

If everyone thought like you, then where is the capacity on the roads to do that?

A single lane of highway can accommodate ~2000 vehicles per lane per hour in one direction. Let very generously say AV's double that. So 4000 vehicles per lane per hour in one direction. 12 Lanes of highway into Toronto (Thats the DVP, QEW, the Spadina Expressway and the Scarborough Expressway have also come to life). In total, that is 48000 people who can come into Toronto based on highway capacity at rush hour.

By contrast, with 15 minute headways, and assuming 2000 people per train, with 7 lines (LKE, Stouffville, Richmond Hill, Barrie, Kitchener, Milton, LKW), those alone provide the capacity for 56,000 people to come into Downtown Toronto by themselves per hour. In the case of GO expansion, there are going to be nearly 52 trains per hour at peak times coming into Union, meaning 104,000 people per hour will be served by those trains. To provide the equivalent highway service would mean constructing nearly 20 new lanes of highway heading into downtown.

Quite simply there isn't enough space in any major city for a majority of the population to abandon public transit and switch over to autonomous vehicles. And for the sake of our cities and urban areas, we cannot allow for your vision of an autonomous vehicle hell hole. Public transit is a necessity to ensure that future cities remain livable for humans. To prevent a Saudi/Dubai like future where highways sprawl everywhere and there is no possibility for a cohesive urban fabric.

Like this picture shows

So as long as humanity has to obey the laws of physics, there will continue to be a need for public transit.

 

[crs1026]

I think that is the problem. Transit agencies and "urbanists" are too stuck in their ways. Transport is changing. Much in the same way that airlines are moving away from large aircraft and hubs to smaller point to point flight, so will AVs bring to land transport. It is just more efficient and provides better customer service.

Transit agencies do innovate, but they can’t change faster than their mandate allows. Imagine the TTC CEO standing up in front of City Council and announcing that the entire bus fleet was about to be traded in on a fleet of AV’s. Even on a good day, Council would at least ask a million pointed questions.....which no one has the answers to yet. And then the Roads General Manager would stand up and ask that their budget be doubled, and they need more time, and, and.....

It’s far too easy to declare “I have a vision” and then accuse anyone who doesn’t move as fast as you like that they are “stuck in their ways”. There is a step or two between “visioning” and “rollout”. It’s not magic, it’s hard planning.

Really, alot of the light rail projects we see under construction are entirely unnecessary in the era of autonomous vehicles.

The same goes for commuter/regional rail and intercity rail. That is why I am against a project like HFR. If I am able to ride in an autonomous vehicle from Toronto to Ottawa, I will never ride the train again because it will have no advantages. I will be able to rest on the journey, get work done and probably arrive directly at my destination. It would also probably be cheaper because AV providers would have very few fixed costs. I could also see it being better for the environment as these AVs would overwhelming be electric whereas VIA will still be running diesel Siemens Chargers (that is if they weren't already shut down).

Putting auto users in an EV -or a Siemens Venture/Charger train will reduce carbon, sure. Now ask your AV to pull over because your child needs to use the bathroom. Or wants fries. AV may help VIA with first mile-last mile, but I don’t see it becoming universal. It will be part of a mix.

- Paul

 

[2transpo]

Quite simply there isn't enough space in any major city for a majority of the population to abandon public transit and switch over to autonomous vehicles.

This arguement doesn't hold true when you consider that the average occupancy of road vehicles will increase with AVs to 2-3 passengers per vehicle.

Also, take into account a smaller city like London or Hamilton. Shutting down the slow, local busses and switching to demand responsive AVs would actually improve mobility in those cities.

Putting auto users in an EV -or a Siemens Venture/Charger train will reduce carbon, sure. Now ask your AV to pull over because your child needs to use the bathroom. Or wants fries. AV may help VIA with first mile-last mile, but I don’t see it becoming universal. It will be part of a mix.

The problem is that when you have EVs, the environmental arguement for rail falls apart, especially going up against diesel trains like VIA will be receiving. An EV will have a better carbon footprint than the train when the train is diesel.

The other concern is VIA's last mile problem could turn into an every mile problem. If I could take an AV to the station, what's stopping me from taking it all the way to Ottawa? If I need anything, I could always pull over at a rest stop like people already do on the highway.

We also need to address that autonomous ride-hailing will also be significantly cheaper than VIA because of lower capital costs, lower fixed operating costs anx lower energy costs. A lot of people will never take the train again when autonomous vehicles are around because the train has nothing to offer that and AV doesn't.

 

[crs1026]

This arguement doesn't hold true when you consider that the average occupancy of road vehicles will increase with AVs to 2-3 passengers per vehicle.

Also, take into account a smaller city like London or Hamilton. Shutting down the slow, local busses and switching to demand responsive AVs would actually improve mobility in those cities.


The problem is that when you have EVs, the environmental arguement for rail falls apart, especially going up against diesel trains like VIA will be receiving. An EV will have a better carbon footprint than the train when the train is diesel.

The other concern is VIA's last mile problem could turn into an every mile problem. If I could take an AV to the station, what's stopping me from taking it all the way to Ottawa? If I need anything, I could always pull over at a rest stop like people already do on the highway.

We also need to address that autonomous ride-hailing will also be significantly cheaper than VIA because of lower capital costs, lower fixed operating costs anx lower energy costs. A lot of people will never take the train again when autonomous vehicles are around because the train has nothing to offer that and AV doesn't.

Your 2-3 companions might object to making stops along the way. Sharing is overrated. What percentage of Uber/Lyft rides today represent sharing? That Toronto-Ottawa traveller will be waiting somewhere for their shared ride companions to turn up, or their companions will already be in the AV (having taken a tedious detour to pick you up) when it turns up at your doorstep. First-mile/last mile will remain a challenge for shared rides. Long Distance AV sharing may require termini of its own.

Figure out the cost of the road infrastructure required before you assert "cheaper". Fixed guideways use less energy, are cheaper to construct and maintain, permit higher speeds. Are we building a land transport system that is tied to AV speed potential? AV's may learn how to drive in snow, but will likely do it slower than today's drivers (who regularly end up in the ditch). AV insurers won't accept the level of risk that the average aggressive driver takes.

For anyone who thinks AV's are going to flood our roads any time soon, I have two words: B737 Max. The umpteen hundred million lines of code in an AV model will have to be examined by regulators and lawyers, line by line. How much is discoverable? How much stored data can a victim's lawyer gain access to after a crash? How many of those lines of code have yet to be scrutinised by a regulator, supported by test data, verified to an appropriate 99.999999% effectiveness level? What jurisdiction has even set standards for how many AV fails are acceptable - one fatality every x million seat-miles? How long will it take the public to accept that standard so it can be codified in law? The B737 Max is now getting rolled out, but I know a ton of people (myself included) who have no intention of flying in one post-COVID until there are a couple of years' crash-free experience to prove that the model is airworthy. And it's a longstanding model, albeit with a few significant design changes, but with lots of attributes that are time-proven. AV's are way behind that certification curve.

AV is coming but it's still a long way from disrupting. EV's will be here much sooner.

- Paul

 

[MrGoose]

This arguement doesn't hold true when you consider that the average occupancy of road vehicles will increase with AVs to 2-3 passengers per vehicle.

Also, take into account a smaller city like London or Hamilton. Shutting down the slow, local busses and switching to demand responsive AVs would actually improve mobility in those cities.

Its not even highway capacity, once you get off the highway, the city streets don't have nearly as much capacity as the highway does. So either way, you're going to get massive traffic jams because the AVs need to stop and people have to get out of the cars.

The other concern is VIA's last mile problem could turn into an every mile problem. If I could take an AV to the station, what's stopping me from taking it all the way to Ottawa? If I need anything, I could always pull over at a rest stop like people already do on the highway.

We also need to address that autonomous ride-hailing will also be significantly cheaper than VIA because of lower capital costs, lower fixed operating costs anx lower energy costs. A lot of people will never take the train again when autonomous vehicles are around because the train has nothing to offer that and AV doesn't.

Lower Capital Costs: How so? The construction of 1 km of highway requires as much if not more earth moving and construction as the equivalent length of rail.

Lower fixed operating costs: Again how so? Either the AVs are driving themselves in which case, they’d do a much better job driving a train since trains operate on fixed guideways and as such their location is much easier to know than a car. Or the car is connected wirelessly to a server farm in which case you’d have to incorporate the cost of running low latency server farms to drive cars.

In fact there are already self driving trains that operate commercially today but autonomous cars have yet to operate commercially en masse. The technology for autonomous train operation is cheaper and easier to build than self driving cars.

Lower Energy Costs: How? If both the av and the train are on electricity, then the train (being more aerodynamic and running on lower friction guideways) will be more efficient at the same speed.

Why would you take a train? Because the train is faster (I highly doubt that avs are going to be traveling on the highway at 300kmph), the train will take you downtown directly instead of having to navigate city streets. The train is cheaper than renting an av for the 4 hours it’ll take to drive to Ottawa or taking the train every week is cheaper than having an Uber AV subscription or the train is cheaper than owning and maintaining your own AV.

 

[lenaitch]

I think the discussion is settled. Hyper Loops and AVs are real-world proven. It's just weak-kneed politicians and misguided SMEs who are standing in the way.

???

 

[afransen]

For anyone who thinks AV's are going to flood our roads any time soon, I have two words: B737 Max. The umpteen hundred million lines of code in an AV model will have to be examined by regulators and lawyers, line by line. How much is discoverable? How much stored data can a victim's lawyer gain access to after a crash? How many of those lines of code have yet to be scrutinised by a regulator, supported by test data, verified to an appropriate 99.999999% effectiveness level? What jurisdiction has even set standards for how many AV fails are acceptable - one fatality every x million seat-miles? How long will it take the public to accept that standard so it can be codified in law? The B737 Max is now getting rolled out, but I know a ton of people (myself included) who have no intention of flying in one post-COVID until there are a couple of years' crash-free experience to prove that the model is airworthy. And it's a longstanding model, albeit with a few significant design changes, but with lots of attributes that are time-proven. AV's are way behind that certification curve.

This is just anti-autonomy FUD. There are several jurisdictions that have essentially legalized autonomous driving, including China, Florida and California. Google is operating driverless taxis in Phoenix for revenue service. So, your argument here is demonstrably untrue.

 

[afransen]

The technology for autonomous train operation is cheaper and easier to build than self driving cars.

This needs to be demonstrated. Costs are completely out of hand for transit projects.

Why would you take a train? Because the train is faster (I highly doubt that avs are going to be traveling on the highway at 300kmph), the train will take you downtown directly instead of having to navigate city streets. The train is cheaper than renting an av for the 4 hours it’ll take to drive to Ottawa or taking the train every week is cheaper than having an Uber AV subscription or the train is cheaper than owning and maintaining your own AV.

We're not getting HSR any time soon, but still, HFR will travel faster than posted speed limits (which may only get raised slightly). And I can't imagine AVs being cheaper than a train ticket. Maybe if you had a car-load of passengers. And sitting in a car for 6 hours is not all that comfortable.

 

[afransen]

I think the discussion is settled. Hyper Loops and AVs are real-world proven. It's just weak-kneed politicians and misguided SMEs who are standing in the way.

???

Hyperloop and AVs are unrelated, let's not conflate.

 

[MrGoose]

This needs to be demonstrated. Costs are completely out of hand for transit projects.

Well, considering that the technology for automatic train operation already exists, works on existing hardware and is already commercially viable, I would think that autonomous train operation is always going to be cheaper. Low-tech solutions will almost always been cheaper than hi-tech solutions. Autonomous cars need LIDAR, machine learning models, and expensive software developers to function. An autonomous train needs a set of along-line sensors to control it. But once that infrastructure is set up, then all the train needs is transmitter and receivers and relatively low-grade hardware.

Even in China, capital of autonomous experimentation, they are thinking about equipping the roads with sensors to facilitate autonomous driving.

 

[afransen]

Autonomous cars need LIDAR, machine learning models, and expensive software developers to function. An autonomous train needs a set of along-line sensors to control it. But once that infrastructure is set up, then all the train needs is transmitter and receivers and relatively low-grade hardware.

Tesla has shown that LIDAR is unnecessary. lol@ 'expensive software developers'. Go tell Google, Facebook, Apple, etc. that they are doomed: too many devs required.

The problem with transit is that it costs way too damn much. Maybe because everything is bespoke, maybe because of the ossified planning process around developing and building them. When Toronto is building subways that cost $6,000/cm, we should think there is a problem. Even just do the math of amortizing how much a subway line costs over the expected 30 year ridership and think about what the capital cost per ride is. Ontario Line is expected to carry 388k passengers/day x 365 days/year x 30 years = 4.3B rides over 30 years. For a capital cost expected to be around $10B. That's around $2.5 capital cost per ride. Never mind operating cost. Probably around $0.25-$0.30 per km, depending on average trip distance. If we want transit to be viable and withstand competition for AVs, we need to get cost under control.

 

[2transpo]

If we want transit to be viable and withstand competition

What if we just said forget about transit and let AVs be our transit? Financially, it would make a lot of sense. Save ourselves the money and grief of building the Ontario Line.

Metrolinx should get in touch with Elon Musk for Tesla AVs or Google for Waymo. Honestly, to be implementing fixed route transit, especially low speed and low capacity light rail is severely anachronistic.

Why would you take a train? Because the train is faster (I highly doubt that avs are going to be traveling on the highway at 300kmph), the train will take you downtown directly instead of having to navigate city streets. The train is cheaper than renting an av for the 4 hours it’ll take to drive to Ottawa or taking the train every week is cheaper than having an Uber AV subscription or the train is cheaper than owning and maintaining your own AV.

The train isn't going to he faster. Factor in door-to-door, not station to station. AV wins hands down every time, especially against slow HFR. HFR is $5 billion for nothing, especially when you account for the fact that AVs will be more comfortable than regular cars since they will have less crash safety requirements as road accidents will be reduced by 90-99%