[lenaitch]

The only thing about AVs that will definitely be good for our cities is the complete elimination of parking, which will definitely help improve the built environment by eliminating the single worst land use.

That conflates AV with ride share. We could all still have an AV in our driveways. Even with ride share, it may, over time, eliminate source and destination (residence, workplace, shopping, etc.) parking requirements but they would still need lots for storage off-peak.
[/QUOTE]

 

[2transpo]

However tbh, I am very worried about autonomous vehicles. I think there is a very real possibility that they will cause the death of our cities and a renewed era of suburban and exurban car-centric sprawl. How can we encourage a built form that is still supportive of active transportation and transit when AVs are around? Autonomous electric car ride-hailing will just be too convenient and desirable.

There is nothing that we can do about this but that is okay. The economics are going to dictate the balance of transport and if it creates new sprawl, oh well. While it creates problems, the benefits of AVs make it worthwhile. The greatest of these benefits are the increased access to transportation for lower income people.

The reality is that the induced demand won't be all that bad because the demand will be met through ride sharing and traffic flow optimization and reduced vehicle size.

One of the great things about AVs are that they will reduce the need for some expensive projects. For example, people want to bring back the Northlander but when autonomous vehicles are available, somebody will be able to book an autonomous shuttle from North Bay to Toronto. The shuttle will get them there comfortably without them needing to drive (great for senior citizens), will be quicker, door to door and run without the need for any taxpayer subsidy. On top of all that, the ride would likely also be cheaper than a train ticket because of low fixed capital costs.

Autonomous vehicles are going to change everything and it is something that we should look forward to. It is something that I used to struggle with since I like trains, but I learned to embrace it and so should other urbanists.

 

[lenaitch]

The way I understand it, full autonomy requires 5G connectivity. That would require a massive build out of the 5G network. Not all of it between population centres would be considered profitable by the carriers w/o public money. We are continually tossing money at them a pleading for them to expand their network into areas they don't want to, with apparently minimal success.

I still remain unconvinced that AVs will make significant penetration outside of urban areas - at least in our lifetime. As mentioned earlier, I await a real-world test on a snow covered county rural road.

Others have different views, and that's okay. I've been wrong about a lot of things in life.

I'm old enough to remember futurists and other experts telling us we'd be zipping around in flying cars, have a mini- nuclear reactor in our basement and have so much free time on our hands from computerization that we wouldn't know what to do with ourselves, all by the turn of the century or earlier.

 

[afransen]

That conflates AV with ride share. We could all still have an AV in our driveways. Even with ride share, it may, over time, eliminate source and destination (residence, workplace, shopping, etc.) parking requirements but they would still need lots for storage off-peak.

[/QUOTE]
It really eliminates the need for mandatory parking provision. And I think most people will reduce or eliminate private ownership of their vehicles as AVs become cheaper.

AV should not need 5G, as all the perception and control happens locally on the vehicle (insane to try otherwise). Network connectivity is just to receive instructions for where to go (for pickups) and maybe diagnostics/remote intervention in exceptions.

 

[sche]

I still remain unconvinced that AVs will make significant penetration outside of urban areas - at least in our lifetime. As mentioned earlier, I await a real-world test on a snow covered county rural road.

Real-world tests of full-self driving software in snow:

Obviously it doesn't work that well in snow currently, but if you followed the progress of Tesla's Full Self-Driving Beta over the past few months the progress is quite astounding. Already works very well in suburban and rural areas in good weather, and still works decently in urban areas or in poor weather.

 

[afransen]

There is nothing that we can do about this but that is okay. The economics are going to dictate the balance of transport and if it creates new sprawl, oh well. While it creates problems, the benefits of AVs make it worthwhile. The greatest of these benefits are the increased access to transportation for lower income people.

The reality is that the induced demand won't be all that bad because the demand will be met through ride sharing and traffic flow optimization and reduced vehicle size.

One of the great things about AVs are that they will reduce the need for some expensive projects. For example, people want to bring back the Northlander but when autonomous vehicles are available, somebody will be able to book an autonomous shuttle from North Bay to Toronto. The shuttle will get them there comfortably without them needing to drive (great for senior citizens), will be quicker, door to door and run without the need for any taxpayer subsidy. On top of all that, the ride would likely also be cheaper than a train ticket because of low fixed capital costs.

Autonomous vehicles are going to change everything and it is something that we should look forward to. It is something that I used to struggle with since I like trains, but I learned to embrace it and so should other urbanists.

Capacity improvements will be minor--the article you linked to showed 7% street capacity improvement--and it will largely be eaten up by increased VMTs caused by deadheading between passengers. Eliminating parking will free up more travel lanes.

 

[afransen]

The way I understand it, full autonomy requires 5G connectivity. That would require a massive build out of the 5G network. Not all of it between population centres would be considered profitable by the carriers w/o public money. We are continually tossing money at them a pleading for them to expand their network into areas they don't want to, with apparently minimal success.

I still remain unconvinced that AVs will make significant penetration outside of urban areas - at least in our lifetime. As mentioned earlier, I await a real-world test on a snow covered county rural road.

Others have different views, and that's okay. I've been wrong about a lot of things in life.

I'm old enough to remember futurists and other experts telling us we'd be zipping around in flying cars, have a mini- nuclear reactor in our basement and have so much free time on our hands from computerization that we wouldn't know what to do with ourselves, all by the turn of the century or earlier.

It doesn't really matter how well it works on rural roads. 99% of trips don't take place on them. It's okay if they don't get solved until decades later (but I doubt it).

 

[lenaitch]

It really eliminates the need for mandatory parking provision. And I think most people will reduce or eliminate private ownership of their vehicles as AVs become cheaper.

AV should not need 5G, as all the perception and control happens locally on the vehicle (insane to try otherwise). Network connectivity is just to receive instructions for where to go (for pickups) and maybe diagnostics/remote intervention in exceptions.
[/QUOTE]

I though for full 'top level' AV connectivity was required so the vehicles could talk to each other. Maybe I'm wrong.

It doesn't really matter how well it works on rural roads. 99% of trips don't take place on them. It's okay if they don't get solved until decades later (but I doubt it).

Well, it does to the people who live there, or the urban folks who want to travel outside of the urban bubble. It was in response to a statement of AVs running between North Bay and Toronto. Hwy 11 can be completely snow packed for large parts of the winter.

 

[afransen]

It really eliminates the need for mandatory parking provision. And I think most people will reduce or eliminate private ownership of their vehicles as AVs become cheaper.

AV should not need 5G, as all the perception and control happens locally on the vehicle (insane to try otherwise). Network connectivity is just to receive instructions for where to go (for pickups) and maybe diagnostics/remote intervention in exceptions. though for full 'top level' AV connectivity was required so the vehicles could talk to each other. Maybe I'm wrong.

Well, it does to the people who live there, or the urban folks who want to travel outside of the urban bubble. It was in response to a statement of AVs running between North Bay and Toronto. Hwy 11 can be completely snow packed for large parts of the winter.

I don't think snow on a highway will be a significant challenge for an AV in a decade.

There will still be manually operated cars for a very long time, which can be used for trips poorly structured roads. Just because there is a dirt track out there an AV can't drive on does not mean they won't be in widespread use in cities.

 

[APTA-2048]

New Flyer unveils the Xcelsior AV™ - North America's first automated transit bus

/CNW/ - (TSX: NFI) New Flyer of America Inc. ("New Flyer"), a subsidiary of NFI Group Inc. ("NFI"), one of the world's leading independent global bus...

www.newswire.ca

Level 4 automated bus highlights New Flyer's dedication to innovation and the strength of its partnership with Robotic Research.

 

[afransen]

It's a pretty big deal, as this should significantly cut the cost of operating buses, and eliminating the driver driver simplifies operations and scheduling. Could also move toward smaller buses (more like 12-20 pax) to maintain acceptable frequency at night, and provide more neighbourhood circulation during the day. Of course, while the cost of autonomy packages (10 lidars, compute, etc) is high, it will make sense to use it for larger vehicles. But I would think within 5 or so years those things would be cheap enough to put on smaller buses.

 

[rbt]

It's a pretty big deal, as this should significantly cut the cost of operating buses, and eliminating the driver driver simplifies operations and scheduling.

Hopefully they don't simply add $500k to the purchase price. I expect it'll be used as a platooning vehicle (follow at close distance a leading bus with a live driver) for many years.

Also worth noting, Robotic Research hardware does not have a California drivers license, so there is a fair amount of R&D to do. It's more than strapping their hardware to a bus and rebuilding models for the vehicles sensor positions, acceleration/deceleration curves, turn radius, etc. Real military deployments is great, but their tolerance for risk is significantly higher than the general public.

 

[lenaitch]

I don't think snow on a highway will be a significant challenge for an AV in a decade.

There will still be manually operated cars for a very long time, which can be used for trips poorly structured roads. Just because there is a dirt track out there an AV can't drive on does not mean they won't be in widespread use in cities.

A reasonable perspective; much more realistic than those who believe they will be in every driveway and warehouse yard in a couple of years.

 

[2transpo]

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.

 

[afransen]

Hopefully they don't simply add $500k to the purchase price. I expect it'll be used as a platooning vehicle (follow at close distance a leading bus with a live driver) for many years.

Also worth noting, Robotic Research hardware does not have a California drivers license, so there is a fair amount of R&D to do. It's more than strapping their hardware to a bus and rebuilding models for the vehicles sensor positions, acceleration/deceleration curves, turn radius, etc. Real military deployments is great, but their tolerance for risk is significantly higher than the general public.

The sensor package is probably on the order of $50k, so not a massive cost increment on a bus. Of course, Tesla says they can do it with $100 worth of cameras and probably a <$500 GPU.