Connected/Automated vehicles and infrastructure in Boston

stefalarchitect said:
Can you or someone else clarify how the private sector is going to pick up some/most of the bill? I don't see how they can make a profit by investing in infrastructure. Seems like the only way we can get infrastructure updates is through infrastructure bills, which are very rare lately, and, like you mentioned, inefficient.
Click to expand...

First off, I don't believe that the benefits to be gained through autonomous driving technology necessitate significant infrastructure spending. The technology to drive fully autonomously on existing roads may not quite be there yet in 2016, but it likely will be there in 202X. And if it isn't quite there for driving in dense cities, it will almost surely be there for driving on highways.

As far as the private sector picking up the bill for infrastructure spending, look no further than New Balance / Boston Landing. If large players in any space believe that Y dollars of infrastructure spending would unlock X dollars of value to their bottom line, and Y<X, then it is in the interests of the private sector entities to make it happen. And given the enormous amount of money that certain entities stand to gain by switching to autonomous driving technology, it becomes likely that Y<X.

And plenty of infrastructure updates do "just happen" when the financial calculus backs them up. Look at the transition to LEDs first in stoplights and now in streetlights. Before long, it'll be hard to find a non-LED light shining on a roadway. This hasn't required any sort of concerted effort, per se, it's just been a process of routine upgrades that make sense financially. I'd imagine that the technology and adjustments needed to necessitate autonomous vehicles would progress similarly.
 
Could someone also provide projections of how the trucking industry will allegedly save so much money to go automated? The elimination of drivers' salaries is easy enough to grasp, and I guess many on this thread are just fine with seeing another whole class of employees get put permanently out of work so we can ship widgets cheaper. (Can you tell I'm not as enthused?)

However, all that software is not going to run itself, especially if we're talking about networked fleet updates on road conditions. There will be server and IT staffers maintaining the servers and the software. Lots of servers, lots of IT staff, and that's just on the truck-specific software; as others have noted, this all might require significant infrastructure upgrade that would have to get paid for.

Will there really be a savings? Has anyone seen actual math on it? I can easily believe that a system like the automated shuttle train at SFO Airport is cheaper without drivers, but that's about as simplistic a fixed-track system as one could imagine, I find it easy to believe the IT support for that is cheaper than some bored-to-death drivers would be. But ... the entire truck fleet, or even a large percentage of it, out on the roads? The needed IT support will be immense, and very expensive.

I'm asking from ignorance here, not from a snarky "I know better" stance.

As for safety, I am quite receptive to the safety improvement possibilities of the shift to automated vehicles, even if it takes longer than optimists' projections (that's my seat of the pants hunch). Many human drivers really suck at the task, and we'd save more than just money by cutting down on the carnage. But the new system will still be human-designed and maintained, and will still fail sometimes. I'm curious where the liability resides under the new system? Answering that to the satisfaction of the insurance and litigation industries could have as much impact on the roll-out time-frame as figuring out the technical software challenges.
 
Boston Landing station is a great example of why private investment in infrastructure is often to the public detriment. The station siting is designed to serve New Balance, without much benefit to the neighborhood. Sites at Market Street and Cambridge Street scoped out much better on public good in the 2007 study they did - particularly because Market gets the 86 bus (and short walk to the 70) and Cambridge gets the 66 and 64 (plus the 57 nearby), while Everett gets zilch. And because New Balance holds all the cards here, they weren't required to do any kind of public input.

So we end up with a station that not only doesn't serve the community well but actively prevents the addition of a properly sited station. That lengthens the trip time for many Metrowest commuters, while serving few of them.* And while they're building it, the entire line suffers from cascading delays - which would have been lessened with a properly planned project. If that's a representative example of what private infrastructure funding looks like, then count me the hell out.

* The combination of the 86 and 70, or the 66 and 64, gives you direct bus connections to Harvard, Central, and Kendall squares that will beat the Red Line on most days. The actual station site has only a poor connection to the 64, the least frequent of those routes.
 
West said:
Could someone also provide projections of how the trucking industry will allegedly save so much money to go automated? The elimination of drivers' salaries is easy enough to grasp, and I guess many on this thread are just fine with seeing another whole class of employees get put permanently out of work so we can ship widgets cheaper. (Can you tell I'm not as enthused?)

However, all that software is not going to run itself, especially if we're talking about networked fleet updates on road conditions. There will be server and IT staffers maintaining the servers and the software. Lots of servers, lots of IT staff, and that's just on the truck-specific software; as others have noted, this all might require significant infrastructure upgrade that would have to get paid for.

Will there really be a savings? Has anyone seen actual math on it? I can easily believe that a system like the automated shuttle train at SFO Airport is cheaper without drivers, but that's about as simplistic a fixed-track system as one could imagine, I find it easy to believe the IT support for that is cheaper than some bored-to-death drivers would be. But ... the entire truck fleet, or even a large percentage of it, out on the roads? The needed IT support will be immense, and very expensive.

I'm asking from ignorance here, not from a snarky "I know better" stance.
Click to expand...

Technology has been replacing human labor at breakneck pace for more than two centuries now. With each new phase of innovation people have lamented how this development would be the end of labor, but we've somehow managed (kept on trucking?). Yes, there have been some losers over the short term, and the economy today is far from perfect, but all things considered we are collectively much better off for it.

As the second part of your question implies, the replacement of one job opens up whole new categories of jobs. For each job that has been replaced by computers (or any other technology) more jobs have opened up working on and with computers (or any other technology). And as technology replaces people and makes any activity cheaper we end up doing more of that activity, greatly dampening the loss of jobs from the replacement of the labor in the first place. The classic example is ATMs. There's a narrative that ATMs killed bank teller employment, but the data shows that the number of tellers employed in America never really dropped with the advent of ATMs. On different threads in this forum we lament the preponderance of banks in our city, but all those banks can afford to stay open because ATMs and other technology brings down their operational costs. Instead of 10 banks employing 300 people we now might get 30 banks employing 300 people: three times the service and no change in employment. And automated banking technology makes everything work way better than it did in the first place. This is obviously a simplified example, but it could apply to trucking as well. Removing the driver makes trucking cheaper not only because you don't have to pay his or her wage, but also because trucks now drive straight through for 24 hours a day without stopping, get into fewer accidents, and always use optimum fuel saving driving behavior (including drafting), among other things. This makes trucking cheaper and more efficient, allowing for more trucks and an increased overall trucking capacity. Now these extra goods and extra trucks require more workers to maintain them, and more workers to distribute their cargo beyond the long-haul highway routes (some displaced long-haul truckers can become short-haul drivers), and more workers to design and maintain the driving and networking technology (as you mentioned), and etc. And on top of that everything we all purchase ("widgets" or food and clothing and energy and medicine and technology and ...) gets slightly more affordable, freeing up money in all of our pockets to pay for other things and employee people in those industries.

As for the calculus of which is cheaper, I haven't seen any thorough analysis on that but I'm sure the cost of server and IT support for the global ATM network is immense too. IT systems tend to scale much cheaper and more easily than physical, human labor. And if autonomous driving is more expensive, then we won't have to worry about the loss of truck driver employment...
 
Last edited:
The EGE said:
Boston Landing station is a great example of why private investment in infrastructure is often to the public detriment. The station siting is designed to serve New Balance, without much benefit to the neighborhood. Sites at Market Street and Cambridge Street scoped out much better on public good in the 2007 study they did - particularly because Market gets the 86 bus (and short walk to the 70) and Cambridge gets the 66 and 64 (plus the 57 nearby), while Everett gets zilch. And because New Balance holds all the cards here, they weren't required to do any kind of public input.

So we end up with a station that not only doesn't serve the community well but actively prevents the addition of a properly sited station. That lengthens the trip time for many Metrowest commuters, while serving few of them.* And while they're building it, the entire line suffers from cascading delays - which would have been lessened with a properly planned project. If that's a representative example of what private infrastructure funding looks like, then count me the hell out.

* The combination of the 86 and 70, or the 66 and 64, gives you direct bus connections to Harvard, Central, and Kendall squares that will beat the Red Line on most days. The actual station site has only a poor connection to the 64, the least frequent of those routes.
Click to expand...

If you are making the case that Boston Landing is a net negative for our transportation network, than that's an argument I haven't heard...

The public powers-that-be could have said no. At the end of the day, it was all subject to their approval. And bus routes can be changed.
 
Boston Landing versus no Boston Landing is probably a break-even, as are many urban infill stops: small inconvenience for most, big gain for a few. But the station is now preventing the substantially better best case scenario from occurring - a properly sited station that serves more local residents and has connections that help a higher number of existing riders. Compared to that scenario - one entirely achievable by 2020 without NB - Boston Landing represents a substantial loss.

RFvnkzK.png


That's what the area looks like. The 70/70A in brown, 64 in pink, 86 in red, 66 in purple, 57 in blue. The Market Street site at left serves one route well and two route okay. The Cambridge Street site at right serves two routes well and one route okay. And yet the chosen site was the Everett Street site at middle that one serves one route at all.

To shift any of those routes but the 64 would be a major issue for many riders - you'd be adding a significant inconvenience to serve thousands of people a day. That's very poor planning and will never happen (and is why despite the MBTA's claims, neither the 57 or the 66 will serve West Station either). The 64 already has a deviation through the area (added in 1998 to serve the Stop & Shop) that's a classic example of poor bus route planning.
 
The EGE said:
Boston Landing versus no Boston Landing is probably a break-even, as are many urban infill stops: small inconvenience for most, big gain for a few. But the station is now preventing the substantially better best case scenario from occurring - a properly sited station that serves more local residents and has connections that help a higher number of existing riders. Compared to that scenario - one entirely achievable by 2020 without NB - Boston Landing represents a substantial loss.


That's what the area looks like. The 70/70A in brown, 64 in pink, 86 in red, 66 in purple, 57 in blue. The Market Street site at left serves one route well and two route okay. The Cambridge Street site at right serves two routes well and one route okay. And yet the chosen site was the Everett Street site at middle that one serves one route at all.

To shift any of those routes but the 64 would be a major issue for many riders - you'd be adding a significant inconvenience to serve thousands of people a day. That's very poor planning and will never happen (and is why despite the MBTA's claims, neither the 57 or the 66 will serve West Station either). The 64 already has a deviation through the area (added in 1998 to serve the Stop & Shop) that's a classic example of poor bus route planning.
Click to expand...

EGE -- this is hypothetical sour grape-aid

No Build would have been the alternative not some ideal station location

The only reason we have a Boston Landing or Sportsville as I prefer -- is that NB decided to invest in what was undervalued real estate -- and NB chose to follow the advice of all of the advocates for public transportation by building a Commuter rail Station.

This station primarily serves as a destination point for people commuting to NB's campus and other development happening as a result -- probably the best model elsewhere in the T system would be Prudential

Any benefit to the commuting pattern of the existing nearby neighborhood is a secondary consideration. As to the future in a decade or so when the NB area is built-up the bus routes can always be re-evaluated -- after all that's one nice thing about buses -- they can move around on the chess board like a knight they aren't pawns forced only to move in a linear fashion limited to where the rails are located.
 
The EGE said:
RFvnkzK.png
Click to expand...

Basically, one end of the Boston Landing station is at a bus stop on the 64, the other end of the station is as close to the 57 as the Cambridge Street site. All you are missing is the 66, which is just a little bit further away than the 57.

The corner of Everett and the pike has the 2 largest office buildings in Allston, and a block away is the largest residential building in Allston. There are multiple under construction residential projects within a couple blocks and another on Penniman Road pulled permits in August. This is all in addition to New Balance.

Also, MassDOT put the kiboche on Market Street; that site would have served New Balance just as well. And the Cambridge street site would have been way too close to West Station, which was always a gleam in the eye of the planners, even if its future was presented as uncertain.
 
CNBC has pretty good article on the 10 industries to be disrupted ("winners and losers") from self-driving cars:
Parking: (loser: demand to fall by half)
Real Estate (fringes win; core loses)[I say: oversimplified*]
Law Enforcement (loser: lost revenues from parking tix)
Insurance (loser: fewer accidents mean lower $ flow)
Legal Professionals (loser: fewer accidents mean lower deal flow)
Hotels (loser: "sleeping car" delivers/retrieves you without a hotel stay)
Media, Entertainment and Online Retail (winner: more screen time)
Package and Food Delivery: (winner: more take-out)
Auto Repair: (loser: fewer repairs)
Auto Manufacturing: (loser: smaller fleet size)

Also a neat statistic:
The U.S. has about 144 billion square feet of total parking, which represents up to one-third of the total real estate in some large cities. Reports estimate self-driving vehicles have the potential to reduce parking space by about 61 billion square feet, which is about the size of Connecticut and Vermont combined. A reduction in the demand for parking can result in reclaiming this valuable real estate for more beneficial social and economic purposes.
Click to expand...

That stat comes from this paper:
Autonomous Vehicles and Commercial Real Estate

*I think it is easy to see that if the burden of a long commute falls that AVs would open up the exurbs to development in exactly the same way the railroads opened up residential land wherever they went (as with train commuting, you'd sleep or work for 2 hours while the car crawls in in the AM and out in the PM), the exurbs are immediate winner for *residential*

I suppose if more talent moves further out, you could argue that at least some jobs would move out because that's where the talent is. Can't that also be argued that the other way: if commuting is painless, it is still the case that the CBD will remain the place that access to maximum talent drawn from all compass points? But back the other way: painless commuting would allow Burlington and Norwood and FiDi to all be roughly equal in claiming to be "the center", that Burlington and Norwood would be the relative winners, even if the core remained top dog.
 
I'm beginning to wonder if self driving cars will actually be "good" for traffic as people often indicate. My understanding is that most of the potential ways self driving cars could reduce traffic actually come from connecting cars which is likely a lot further out if that happens at all (the security concerns are what I see stopping this.) But I see several places where they could increase traffic:


  • As Arlington's already alluded, lowering the burden of driving might make people elect to ride in automated vehicle than those who were willing to drive and people might be willing to ride for farther than they were willing to drive. This could increase the amount of single occupancy vehicle commuters and increase the distance of their commutes. The effect this would have really depends on how much people are willing to take longer commutes with automated cars.
  • If we transition to more ridesharing apps, the massive amount deadheading would significantly increase vehicle miles traveled especially in more developed areas.
  • People sending their cars out to circle blocks, go home, or go far enough away to find cheap parking in order avoid parking goes completely against how our cities are currently designed. Any road that currently already has congestion both ways at peak times will become almost unusable. Also turning on and off certain streets will become much difficult as peak roads are all congested in both ways during every peak. Also imagine the hell of everyone downtown attempting to summon automated cars at the end of the day.
  • Stemming from the last two points, we would face issues with too many people trying to get into cars curbside.

So unless automated vehicles are much better at reducing than I predict, I foresee them creating a lot of traffic and infrastructure issues (not to mention environmental). Here's a few policies I can think of that would mitigate this:


  • A significant per mile fee on empty vehicles (this could be hard to enforce)
  • Significant encouragements to multi occupancy vehicles
  • Ban or disincentivize door-to-door service (obvious exemptions for the disabled) by creating designated lots/areas for pick-ups and drop-offs with and banning or fining pick-ups and drop-offs outside these areas. This need not be so strict outside developed areas.
  • Congestion fees with fines based on miles traveled in the congestion zone and entries into the zone.
 
Someone explain to me how insurance loses out when the streets are safer. They have to pay out less, and everyone is more or less required to have auto insurance. And insurance companies are actively helping the development of self driving cars.
 
DominusNovus said:
Someone explain to me how insurance loses out when the streets are safer. They have to pay out less, and everyone is more or less required to have auto insurance. And insurance companies are actively helping the development of self driving cars.
Click to expand...

Insurance often makes most of it's money on the difference between taking in premium (payments) and having to shell it back out in claims. They've got giant investment arms they stick that money in and profit on, just like how your bank makes money on your savings. Without that, they'd often be a low margin business. Quite a few insurers lose money on underwriting entirely and make all their profit from that investing.

Lower premiums from fewer accidents means less cash flow and less on hand to use for those investments.

Competition will also likely squeeze out underwriting margins, as lower risk means it's easier to price a policy and be sure you'll make money on it, whereas a more volatile situation demands much higher skill/talent because there's more potential to lose your shirt betting wrong.

--------

The other wrinkle for insurers, is that automated cars might largely get rid of the private auto insurance market.

After all, if your automated car crashes for something other than maintenance you failed to do, it's not your fault, it's the fault of Tesla, GM, whoever built/programmed the thing.

I see it as likely that insurance for such vehicles will be through the manufacturer, and those manufacturers will either be large enough to self-insure or to negotiate a much better deal with a major insurer than a private citizen is likely to be able to.
 
Has anyone seen Elon Musk's most recent and most insane (quite frankly most idiotic) idea? It's essentially an automated highway/road system underneath/between cities (see: subway system), but you're in your own car.

The only thing I like about this venture is the name. (The Boring Company. "Boring. Its what we do.")

We've gone over this topic ad nauseam, but adding more lanes is not the answer. He refutes this by saying we'd be working in 3 dimensions, so in other words, "keep adding lanes by boring more tunnels underneath the ones we just built, and if those fill up, just keep adding more." Also, the price associated with boring miles and miles of tunnels is ridiculous. He responds to this by saying the smaller diameter as a result of decreased ventilation and automated driving leading to more narrow lanes decreases the cost up to 10x and the construction is also accelerated. I think he's exaggerating a little there, and if you apply his "just add more tunnels if they fill up" logic, the price of adding another tunnel would easily double, if not triple, the price of simply adding or upgrading a new subway/transit line. I don't know, maybe he'll prove us all wrong, but at the rate he's working on this, treating it as a hobby, I highly doubt it. Or maybe it's too soon, and this might be a system we use in 50/75 years. We'll see...
 
I'm pretty sure no municipality that matters will actually go along with Musk's tunnel plan, thank goodness.
 
stefalarchitect said:
Has anyone seen Elon Musk's most recent and most insane (quite frankly most idiotic) idea? It's essentially an automated highway/road system underneath/between cities (see: subway system), but you're in your own car.

The only thing I like about this venture is the name. (The Boring Company. "Boring. Its what we do.")

We've gone over this topic ad nauseam, but adding more lanes is not the answer. He refutes this by saying we'd be working in 3 dimensions, so in other words, "keep adding lanes by boring more tunnels underneath the ones we just built, and if those fill up, just keep adding more." Also, the price associated with boring miles and miles of tunnels is ridiculous. He responds to this by saying the smaller diameter as a result of decreased ventilation and automated driving leading to more narrow lanes decreases the cost up to 10x and the construction is also accelerated. I think he's exaggerating a little there, and if you apply his "just add more tunnels if they fill up" logic, the price of adding another tunnel would easily double, if not triple, the price of simply adding or upgrading a new subway/transit line. I don't know, maybe he'll prove us all wrong, but at the rate he's working on this, treating it as a hobby, I highly doubt it. Or maybe it's too soon, and this might be a system we use in 50/75 years. We'll see...
Click to expand...

Musk seems to be trying to address the problems that:

  • Boring tunnels is too expensive
  • Our main traffic arteries don't have enough throughput

While doing this, he is ignoring the bigger problems that:

  • Infrastructure connecting tunnels to the surface is too expensive
  • Everyone likes to get on/off major arteries at the same places and the roads they are coming from/going to have a much lower throughput.

I honestly view this as a sign that Musk is overextending himself and hasn't bothered to listen to a traffic planner, and/or he has been told how automated vehicles could cause radically different infrastructure demands than the ones our cities are built for and is desperately looking for some solution.
 
Arlington said:
CNBC has pretty good article on the 10 industries to be disrupted ("winners and losers") from self-driving cars:
Parking: (loser: demand to fall by half)
Real Estate (fringes win; core loses)[I say: oversimplified*]
Law Enforcement (loser: lost revenues from parking tix)
Insurance (loser: fewer accidents mean lower $ flow)
Legal Professionals (loser: fewer accidents mean lower deal flow)
Hotels (loser: "sleeping car" delivers/retrieves you without a hotel stay)
Media, Entertainment and Online Retail (winner: more screen time)
Package and Food Delivery: (winner: more take-out)
Auto Repair: (loser: fewer repairs)
Auto Manufacturing: (loser: smaller fleet size)
Click to expand...

Another big loser will be people on the organ transplant waiting list.

stefalarchitect said:
We've gone over this topic ad nauseam, but adding more lanes is not the answer.
Click to expand...

Adding more lanes can be an answer if they're toll lanes.

whittle said:
I'm beginning to wonder if self driving cars will actually be "good" for traffic as people often indicate. My understanding is that most of the potential ways self driving cars could reduce traffic actually come from connecting cars which is likely a lot further out if that happens at all (the security concerns are what I see stopping this.) But I see several places where they could increase traffic:


  • As Arlington's already alluded, lowering the burden of driving might make people elect to ride in automated vehicle than those who were willing to drive and people might be willing to ride for farther than they were willing to drive. This could increase the amount of single occupancy vehicle commuters and increase the distance of their commutes. The effect this would have really depends on how much people are willing to take longer commutes with automated cars.
  • If we transition to more ridesharing apps, the massive amount deadheading would significantly increase vehicle miles traveled especially in more developed areas.
  • People sending their cars out to circle blocks, go home, or go far enough away to find cheap parking in order avoid parking goes completely against how our cities are currently designed. Any road that currently already has congestion both ways at peak times will become almost unusable. Also turning on and off certain streets will become much difficult as peak roads are all congested in both ways during every peak. Also imagine the hell of everyone downtown attempting to summon automated cars at the end of the day.
  • Stemming from the last two points, we would face issues with too many people trying to get into cars curbside.

So unless automated vehicles are much better at reducing than I predict, I foresee them creating a lot of traffic and infrastructure issues (not to mention environmental). Here's a few policies I can think of that would mitigate this:


  • A significant per mile fee on empty vehicles (this could be hard to enforce)
  • Significant encouragements to multi occupancy vehicles
  • Ban or disincentivize door-to-door service (obvious exemptions for the disabled) by creating designated lots/areas for pick-ups and drop-offs with and banning or fining pick-ups and drop-offs outside these areas. This need not be so strict outside developed areas.
  • Congestion fees with fines based on miles traveled in the congestion zone and entries into the zone.
Click to expand...

If self-driving cars actually become a thing, they will have to be taxed fairly heavily. It's pretty simple economics: if it becomes really cheap to have a car drive around the streets, then we'll get more cars driving around the streets. That will worsen traffic, assuming that the nature of road infrastructure doesn't change equally rapidly to keep up (which we all know won't happen). We'll then need to bring the price of driving a car around the streets back up, and we can do this with a tax.
 
whittle said:
[*]Ban or disincentivize door-to-door service (obvious exemptions for the disabled) by creating designated lots/areas for pick-ups and drop-offs with and banning or fining pick-ups and drop-offs outside these areas. This need not be so strict outside developed areas.
Click to expand...

Pardon my sounding like someone from the 20th century, but this sounds remarkably like public transportation (or private versions thereof, such as park-and-ride) to me.
 
Coyote137 said:
Pardon my sounding like someone from the 20th century, but this sounds remarkably like public transportation (or private versions thereof, such as park-and-ride) to me.
Click to expand...

That's valid.

My idea was for some policy that might encourage ridesharing services to look like some sort of hybrid of Bridj and slugging.
 
You must log in or register to reply here.

Back
Top