Connected/Automated vehicles and infrastructure in Boston

Scipio

Active Member
Joined
Feb 22, 2012
Messages
482
Reaction score
1
Lots of exciting things are happening with vehicle automation and vehicle to vehicle/infrastructure/bike-ped communication, but mostly in Michigan and the Bay Area. MIT, Volpe and some other locals have been doing good work, but there has been no real ecosystem here. Two developments over the past month have been interesting and could lead to a stronger research scene and testing and experimentation on our roads:

Toyota lab coming to MIT:

http://news.mit.edu/2015/csail-toyota-25-million-research-center-autonomous-cars-0904

It's a $25M commitment paired with another lab at Stanford. This has led to some legislative pressure to make sure that testing automated vehicles on public roads is legal in Massachusetts.

Audi working with Somerville:

http://autotechnews.net/2015/11/20/...r-on-self-parking-infrastructure-in-the-area/

It sounds like they are working on making it possible for Audis to valet park without a driver at Assembly Square. The other part sounds like vehicle to infrastructure communication where traffic light timing information is broadcast to Audis. Somerville and Audi have signed an agreement, hopefully Somerville was smart enough to push for open standards and communications formats.

---

This should be interesting to follow, especially given the contrast between urban Boston as a testing environment versus the suburbs of the Bay Area, Austin, and Ann Arbor.
 
V2V is a near term dead end. Autonomous cars need to be able to drive with their cameras and obey the rules of the road without additional infrastructure.

Only down the road when autonomous vehicles are already a significant percentage of traffic could V2V be useful for coordinating between vehicles for cooperation like drafting and closer spacing on the highways.

But yes it would be good to make sure that laws are updated to accommodate autonomous vehicles.
 
Lots of exciting things are happening with vehicle automation

Toyota lab coming to MIT:

http://news.mit.edu/2015/csail-toyota-25-million-research-center-autonomous-cars-0904

It's a $25M commitment paired with another lab at Stanford.

From the MIT story -- its looks to be a powerful team:

Led by CSAIL director Daniela Rus, the new center will focus on developing advanced decision-making algorithms and systems that allow vehicles to perceive and navigate their surroundings safely, without human input.

“We are excited to mark the start of our partnership with Toyota, and hopefully the beginning of the end for traffic fatalities,” says Rus, the Andrew and Erna Viterbi Professor in MIT’s Department of Electrical Engineering and Computer Science. “Together we have developed some research directions that have the potential to be game-changers in the field, and we look forward to working closely with Toyota and Stanford to make them real.”

Toyota’s larger academic collaboration with MIT and Stanford will be coordinated by Gill Pratt PhD ’89, a former MIT professor who most recently served as program director at the Defense Advanced Research Project Agency (DARPA)’s Defense Sciences Office.

Daniela Rus took over the robotics lead at MIT when Rod Brooks left to found Rethink Robotics

Gill Pratt was the key player behind the walking machines that have sprinted forward both at MIT [Cheeta]
https://www.youtube.com/watch?v=_luhn7TLfWU

and the complete menagerie [Atlas, BigDog and friends] created at the former Boston Dynamics [now owned by Google or Alphabet]

https://www.youtube.com/watch?v=-e9QzIkP5qI
 
/\/\/\ Boston Dynamics took a hit recently after losing a military contract but Google had other plans for them anyways.

There's a Globe piece today about automated vehicles in Massachusetts in general. The general tone is that academia and government here have missed the opportunity to lead, but that there's still some good activity. There's a Singapore startup with a Cambridge office joining Toyota, MIT professors saying that MIT has already missed the (automated) boat, and MassDOT begins automated vehicle regulation discussions next month.

http://www.betaboston.com/news/2016/01/21/can-mass-catch-up-on-self-driving-cars/?p1=Main_Headline
 
Google is planning on selling Boston Dynamics with Toyota or Amazon being likely buyers.

http://www.bloomberg.com/news/artic...put-boston-dynamics-robotics-unit-up-for-sale

There has been a lot of speculation about this over the last couple days, but to sum it up

  1. It has been strongly speculated BD and Google have not been a culture fit. West Coast google with mostly a younger software engineering base compared to BD with a 67 year old CEO and mature mechanical engineering talent.
  2. Andy Rubin left Google.
  3. Google own's Schaft which uses a cheaper and more cost-effective movement technology than BD's hydraulics.
  4. Google likely has already pillaged from BD all the good ideas and people it wants to keep going forward.
  5. Google wants nothing to do with building Military AI which is what BD was primarily about before acquisition.
  6. BD's value to Google seems to be only going down at this point. Selling them and recouping as much of the acquisition cost is worth it. Google's gotten what they want from BD at this point (tech + people).
 
There has been a lot of speculation about this over the last couple days, but to sum it up

  1. It has been strongly speculated BD and Google have not been a culture fit. West Coast google with mostly a younger software engineering base compared to BD with a 67 year old CEO and mature mechanical engineering talent.
  2. Andy Rubin left Google.
  3. Google own's Schaft which uses a cheaper and more cost-effective movement technology than BD's hydraulics.
  4. Google likely has already pillaged from BD all the good ideas and people it wants to keep going forward.
  5. Google wants nothing to do with building Military AI which is what BD was primarily about before acquisition.
  6. BD's value to Google seems to be only going down at this point. Selling them and recouping as much of the acquisition cost is worth it. Google's gotten what they want from BD at this point (tech + people).

Mediapal -- you've got many of the pieces, but your logic path is almost inverted

BD has been approaching rapidly to the holy grail of ambulatory robotics -- how to walk naturally -- mostly funded by DARPA and Gill Pratt

The key, quite obviously is to make as many decisions locally and as close to where the physics meets the road -- Google doesn't like this as it goes against their model of computation and storage

The other dispute is mostly about BD people who have the expertise not wanting to be told how to do what they are best at doing by people in CA or Tokyo

So while Google probably got access to a lot of ideas most probably were ignored -- Amazon or the Toyota -- MIT -Stanford Autonomous Auto entity are much better fit

on the military front the real front-line folks are still trying to find a real mission for these big robots -- plenty of opportunities for robots you can carry in one hand -- not too many for robots that can carry you in one hand
 
If we are still in the urban counter-insurgency business, then humanoid robots that can open doors and clear a house without worrying about IEDs, armed insurgents and can hear the heartbeats of scared children hiding in closets are better than just a small flying drone with a camera viewed on a 6" smartphone screen.

I could see DoD money going towards that type of development for a while.

The next question is whether I can rent a humanoid robot at Home Depot and can it build my shed?
 
If we are still in the urban counter-insurgency business, then humanoid robots that can open doors and clear a house without worrying about IEDs, armed insurgents and can hear the heartbeats of scared children hiding in closets are better than just a small flying drone with a camera viewed on a 6" smartphone screen.

I could see DoD money going towards that type of development for a while.

The next question is whether I can rent a humanoid robot at Home Depot and can it build my shed?

Tangent -- today's humanoid robots aren't yet the match for a well trained human -- where they can have significant role is in traditional 3Ds of robotics:
1) too Dangerous -- such as Fukushima -- but they weren't quite there yet
2) too Dull -- repetitive boring tasks -- don't really need the humanoid for most of these the Baxter and cousins of generalized but still quite task-centric are a lot cheaper and do the job quite well
IMGP4261-1292x580.jpg

3) too Dirty -- sort of self explanatory
 
Onrush of driverless cars slows a proposal to add Green Street Garage parking spaces


“Even two years ago I’d say we need this, but now I’m in a position of recognizing that autonomous driving is coming very rapidly,” councillor Nadeem Mazen said. Self-driving cars have the potential to “make traffic and parking a totally different issue as late as five years from now, but not much later. We’re in a place where we could conceivably study this [garage expansion], budget for this and achieve this – and find we don’t need it. I think the evidence is quite strong that will happen.”

I get that planning infrastructure like a garage is becoming more difficult, but aren't autonomous vehicles still going to need midday and overnight layover space? I'd rather have them out of sight and out of mind than racking up VMT heading out to the burbs during times of low demand.
 
Last edited:
Onrush of driverless cars slows a proposal to add Green Street Garage parking spaces




I get that planning infrastructure like a garage is becoming more difficult, but aren't autonomous vehicles still going to need midday and overnight layover space? I'd rather have them out of sight and out of mind than racking up VMT heading out to the burbs during times of low demand.

Scipio -- I think the point is that a garage for automated vehicles could be very different than a garage that has to accommodate only or mostly human driven vehicles

If you remember the hand held game involving sliding blocks around to change the order of numbers or letters --- that's the basic idea

Especially for all day or all night parking the packing factor could be substantial -- obviously this works less well for places with a lot of turnover
 
Self driving cars would substantially lower the cost of services like Uber or Lyft. This would make it more likely for people to forgo owning their own cars and just use car sharing services instead which would dramatically reduce the demand for parking.

Here is a boston.com pic that shows how much more efficient garages with self driving cars will be.

RECROPArrowstreet_driverless_cars_Phase1.jpg
 
Three Globe articles last week:

nuTonomy, the Cambridge/Singapore startup, is starting a very small test of self driving taxis in Singapore: https://www.bostonglobe.com/busines...riving-cabs/LSvfTyf5W2KIx1tYeN0FtL/story.html

The Toyota lab is using a highly instrumented Lexus to collect information about Cambridge roads, presumably for automated vehicle testing: https://www.bostonglobe.com/busines...driving-car/TG3T3HuFhJmMT3WnaiShzH/story.html

And Bridj is moving into local freight using small sidewalk drones:
http://www.bostonglobe.com/business...g-boxes-bus/6vjAmqSiCRzpdTufTJ3SHO/story.html
 
The Toyota lab is using a highly instrumented Lexus to collect information about Cambridge roads, presumably for automated vehicle testing: https://www.bostonglobe.com/busines...driving-car/TG3T3HuFhJmMT3WnaiShzH/story.html

Makes sense - I was talking with one of the Google self driving car engineers a year or two ago at IO, and he pretty much said that the had 90% of driving in the country down - highways, west coast cities like LA, were no problem. Their only big obstacle were cities like Boston/Cambridge where nothing made sense and the streets were very compact/dense. He went on as much to say that basically it was the last remaining big problem they were facing, and that we were, in fact the benchmark city to get though :)
 
http://news.sky.com/story/lyft-president-john-zimmer-predicts-rise-of-driverless-cars-10584580

Lyft president predicts that all of its cars will be driver-less in five years. If that holds up than maybe Chiofaro and other owners of garages could just tare them down without having to worry about replacing spots. In the future driverless cars could basically end the need for downtown parking, cars would be in use for more of the day and when not in use they could stay parked in the outskirts of the city.

I know the timeline is very ambitious but the possibilities gained from driverless cars would be endless.
 
Makes sense - I was talking with one of the Google self driving car engineers a year or two ago at IO, and he pretty much said that the had 90% of driving in the country down - highways, west coast cities like LA, were no problem. Their only big obstacle were cities like Boston/Cambridge where nothing made sense and the streets were very compact/dense. He went on as much to say that basically it was the last remaining big problem they were facing, and that we were, in fact the benchmark city to get though :)

I think Pittsburgh is a good testing-ground for cities with no known logic to their streets PLUS they have tons of steep hills and the most bridges of any city in the US (maybe the world...I forget the propaganda). In any case, if Uber and CMU are successful there, I think Boston is probably a bit less difficult.
 
I think Pittsburgh is a good testing-ground for cities with no known logic to their streets PLUS they have tons of steep hills and the most bridges of any city in the US (maybe the world...I forget the propaganda). In any case, if Uber and CMU are successful there, I think Boston is probably a bit less difficult.

Why would hills and bridges pose any difficulty for self-driving cars? Bridges are basically just perfectly straight divided highways.

What will be difficult for self-driving cars are unusual traffic patterns, narrow and windy congested roads, tricky merges, turns across traffic, and unpredictable behavior by the other cars on the road. In other words... Boston driving.
 
Why would hills and bridges pose any difficulty for self-driving cars? Bridges are basically just perfectly straight divided highways.

What will be difficult for self-driving cars are unusual traffic patterns, narrow and windy congested roads, tricky merges, turns across traffic, and unpredictable behavior by the other cars on the road. In other words... Boston driving.

Not sure about Boston, but I am fairly certain that a self driving car was doing pretty well on Cambridge City streets about ten years ago...

Still I think we are ten years into a twenty year transportation revolution. I wouldn't expect my next car to be autonomous, but maybe the one after that.

I think the bad weather... You know like the kind of weather humans should be smart enough not to drive in... is probably the next big test for autonomous vehicles considering a lot of companies have been focusing on testing in warmer areas, I would guess not a lot of Winter miles have been driven. Maybe not even to successfully drive in, but to successfully tell the passenger that they should go back inside and drink a hot chocolate because the roads are covered with snow.

Maybe we shouldn't be trying to completely replicate human driving, considering how bad we are at it.
 
It's too bad they're putting in the automatic tolling at the sumner tunnel entrance, because I would have loved to see a self-driving car try to get through the toll booths to the tunnel.

It would be hard enough coming from 1A - but trying to enter from the city-street intersection that's appended to the 90-deg turn in the highway right before the booths - not gonna happen.

I think it would probably just sit there patiently waiting it's turn - until the end of time. Or maybe under the right conditions you'd see a few emergent donut patterns develop.

That place is just a giant game of slow motion chicken, and its a perfect example of what can be so funky about Boston. Or maybe it's more like an unstructured multilateral negotiation, conducted only with body language.

Come to think of it, they must have a bit of a challenge on their hands making that set-up work with 'high speed' tolling for plain-old human drivers next year...
 

Back
Top