News   Mar 28, 2024
 79     0 
News   Mar 28, 2024
 168     0 
News   Mar 28, 2024
 245     0 

Self-Driving Vehicles/Autonomous Vehicle Technology

Mercator

New Member
Member Bio
Joined
Jun 2, 2013
Messages
53
Reaction score
7
Discussing Toronto's approach to the introduction and integration of self driving vehicle technologies.
 
I dont believe such a thread has been created to date, so forgive me if this is a duplicate.

I came across this discussion paper that was prepared by David Ticoll (Munk-U of T) which is an interesting analysis of the potential impacts and outcomes of automated vehicle technology on Toronto's roadways.

http://www1.toronto.ca/City Of Toro...ons/Reports/Driving Changes (Ticoll 2015).pdf
Discussion paper David Ticoll Distinguished Research Fellow Innovation Policy Lab Munk School of Global Affairs University of Toronto October 15, 2015

To those that already follow this technology closely, Ticoll does not discuss anything particularly new.

My hope is that, given that Ontario now has a relatively easy process for getting an automated vehicle license for 2016, that Ontario - and perhaps the GTA - can become a a known testing ground and development centre for this technology.
 
Excellent topic. Not sure Toronto is really all that bothered to even have regulatory oversight of any kind on this, given that this falls in the province's (no pun intended) wheelhouse. However, the implications for Toronto are huge. I see lower parking revenues. And maybe even worse congestion. Nothing stopping someone from telling their car to circle the block for half a day. Cheaper than parking.
 
I want to see self-driving buses and streetcars and on-street LRTs. Obviously, it would be a bit more expensive than a self-driving car, and there are issues to solve (how does a self-driving bus or streetcar know when to stop for passengers? Pressure at the stop, a button, visual sight?), but really, that would save the TTC a metric ton of money on the long run.
 
Here's some pertinent information, which I posted in another forum.

Useful information about NTHSA self driving capability levels

Level 0 -- Your old manual-shift car

Level 1 -- Your newer car with cruise control.

Level 2 -- That fancy car with automatic lanekeeping and adaptive cruise.

Level 3 -- You can safely text behind the wheel, but must intervene in an alarm.

Level 4 -- It can self-valet empty. It also can drive your kids unaccompanied to hockey practice in the middle of a snowstorm.

Tesla Autopilot is nearly Level 3; although legally it must be treated as Level 2 with full attention mandatory, and working/surfing/texting still not allowed. Eventually we may reach a point where Level 3 self-driving vehicle drivers are legally allowed to do light (interruptible) work like reading, watching movies, texting, etc, but must mandatorily intervene, say, within 10 seconds of an alarm (seatshaker, wheelshaker, klaxon alarm, etc).

_______________

Level 4 self-driving cars is going to be a big, wonderful, beautiful, very fancy Pandora's Box, with both rainbows/unicorns and skull/crossbones beaming out of it simultaneously, both utopia and dystopia. You will yank the giftwrap off the Pandora Box, suddenly causing it to pop Jack-In-A-Box style in a beautifully kablooey pop in a big shower of confetti/glitter -- figuratively speaking.

  • Efficiency -- With the prospect of empty vehicles going home to do tasks for kids or spouses, this could be a traffic disaster for freeways. Legislation may be needed if people abuse the privelage of sending empty vehicles dozens of miles.

  • Moral -- Cars that are faced with an unavoidable fatality decision are going to decide whether to save the occupants or pedestrians. Picture the scenario of a baby stroller suddenly running in front of the car at the last second, from behind roadside newspaper boxes (unavoidably unseen by the car's sensors until too late; now a fatality has to happen). Car must instantly decide to crash into baby stroller OR suddenly veer into a parked car/lamppost 1 meter to the side. Legally solve this. Now consider the sole occupant of car is your child being soccermomed unaccompanied to school. Whose life goes? Whose Responsibility? Legislative? Insurance? Etc.

  • Manned/Unmanned interactions Unmanned vehicles interacting with manned vehicles (bicyclists, drivers, buses, ambulances, police cars). How can a police car pull over an empty vehicle for an expired plate? Will governments be comfortable legally allowing empty vehicles? Will police be? Etc.

  • Regulatory -- What you're allowed to do and what you're not allowed to do. People without a driver license stepping into a car? Drunk people stepping into the backseat of an empty self-driving car? How old must be children to go unaccompanied in a driverless car? Mailicious passengers trying to damage a self-driving taxi into causing an accident? Are you allowed to sleep for 8 hours in the bed at the back of a self-driving RV, or truck cab of a self-driving truck? What about self driving public transit (uber scale? carpool scale? minibus scale? large bus scale?) Etc. Etc. Etc.

  • Robustness -- How many redundant sensors and cameras must a self-driving car have (e.g. safely function at damage/loss to 25% of sensors? 50% of sensors? 75% of sensors?), so they don't cause accidents when flying road debris damages a camera. This also affects regulatory and insurance, and creates ideas of futuristic safety testing regimens, to ensure they can survive major damage and still safely recover.

  • Insurance -- What the insurance companies are willing to let you do with a Level 4 driverless car. Including all the above.

  • Safety -- Can a level 4 self-driving car safely drive in the middle of a blinding record rainstorm or major snowstorm blizzard, while carrying children that don't know what to do in an emergency? Even Google Car is currently unable to drive reliably in a rainstorm at this time. Level 4 chauffers (like unmanned Uber) isn't going to be legal until you solve this.

  • Security -- Must be upped massively. Hackers. It's already happened. Hackers remotely kill a jeep causing the car to almost park itself on a freeway! And hackers have already blinded driverless cars. Laser pointer tricks a driverless car.

  • Cost -- The cost of solving all the above, factored into your car's sticker price, your government taxes, and your monthly bills (including insurance). It may be so expensive that most carowners will give up carownership, and just hail a neighbour's empty unused car, coming over to your house Uber-style (as a result, conveniently paying a part of that neighbour's car bill!).
We will see a hell of a lot of Level 3 soon (Tesla Autopilot is already almost Level 3).

We will even see Level 4 in designated areas (e.g. campuses, special lanes, designated roads, etc), maybe not too long after, but would not include the necessary intelligence to self-taxi people yet.

But the full legally allowed Level 4 self-chauffering freedom will build up like the Big One (the earthquake) for a VERY LONG TIME, and then go pop in a spetacular fancy Pandora Box of wonderfulness.

When will full Level 4 freedom occur?

Predictions vary widely, but is completely possible because of difficult thorny steps, it may not be until the very rough neighborhood of 2040s/2050s/2060s/2070s before we finally solve ALL THE ABOVE BULLETS to drive your children in the middle of a snowstorm, for 100% complete freedom on all Canada roads -- i.e. wide-open public roads (rather than private campuses or special lanes) -- and the old manual-drive cars on the roads fall apart and being retired -- before we see GTHA roads full of full Level 4 freedom self-driving cars. It is really a BIG step, because of all the above.

Looking forward to it! Would like to sleep in the back of a car while going to Ottawa at night, whenever the full Toronto-Ottawa highspeed trains are booked solid during holidays.
 
Last edited:
The issue of bad weather is obviously especially pertinent in our climate, and I think it is no accident (ha!) that almost all the autonomous vehicle work is being done in sunny California. Rain is not nearly as bad as snow, which not only reduces vision overall, but can especially obliterate a lot of the visual cues as to where the road even is. I'm guessing that milder climes will have legal self-driving cars long before they get approved for places with winters as snowy as ours.
 
Also, there are often two approximate schools of thoughts:

A utopian version of the self-driving future is the ability to press a button and have your favourite car drive up (Uber-style or rental car self-valetting to you). Efficiencies afforded by one car being used by many people, and perhaps thousands functioning in highly efficient UberPOOL style operation, so frequent that less than 1 minute passes before a car picks you up and doesn't even make too many diversions to drop off existing passengers (Because of sheer number of dropoffs exactly on your carpool route, thanks to highly efficient route-optimizations). Freeing up road capacity all around, because of more people sharing fewer cars. And when you feel like driving yourself, a rented Tesla will drive right up to your driveway, for your weekend fun. And these drivereless cars will chauffer you to the nearest rapid transit route. Suburban bus routes will go extinct; being replaced by highly efficient, faster, more convenient electric "Dial-A-Bus" / "UberPOOL" style systems with highly efficient route-optimization.

A dystypian version of the self-driving future is everybody will take them, sole-occupant, all at peak, and still clog the roads. And make up for the longer commute by sleeping/working/playing in the cars. And made worse by further road clogging, by people sending empty cars home by 30+ kilometers to do duties for other family members (rather than treating cars like a fleet of Ubers, Carshares, or Rentals). Nobody using them as public transit connectors, nor not using them uberPOOL style because people like surfing/reading/sleeping/etc alone in a car (longer commute, but people remain productive). More cars on the road than ever before, more congestion, even worse.

_______

The reality will probably be somewhere in between, needing some regulations -- e.g. to discourage (or make expensive) people sending their own cars empty for long distances driving home to serve their spouses and kids. Road-clogging moves like these. Instead, it would be far cheaper or profitable for the distant car to do selftaxi duty (when it is not in use) in its area (ala like a Uber or become a temporary rental while you're at work), and have the spouse and kids instead hail a car closer to their own area.

New businesses and economics will pop up. Any problems with cleanliness of the robo-chauffering fleet of cars -- would be done via guarantees, and the virtue of being able to self-valet to cleaning services if incidents occur, so the car's automatically already clean, etc. New businesses would pop up; car tuneup services and cleaning services for unattended/unaccompanied cars; tuneups and cleaning businesses can run 24/7 a day, winter tire changes, rustproofing, and not need you to take any time off. You don't need any appointments anymore, you just send the car away at 3:00am to do its own maintenance! And if it breaks down, a robo-towtruck will do it for you. Just go to work, here's a robo-Uber to bring the rest of your way to work, let them handle your broken-down car for you, sir!

How will it affect public transit? Rapid transit corridors will probably boom (e.g. subways, commuter trains, high-traffic bus/LRTs, high speed trains) with self-driving minibuses/Ubers/transit/taxis/etc becoming one of the many last-mile connecting options, while low-traffic suburban buses may die (e.g. UberPOOL style services may be more profitable for lower fares per person, and TTC/GO may run UberPOOL style dial-a-bus services much more efficiently than a fixed-route suburban bus, for example -- assuming they keep dynamic pickup/dropoff routes as ultraefficient as possible, and number of occupants per vehicle low -- then it can probably beat the transit economics and transit performance performance of a low-usage suburban bus route).

For all the above reasons, generally, any high-quality rapid transit (subways, LRTs, etc) being built today (e.g. Hurontario LRT, Crosstown LRT, Hamilton LRT), will generally fit extremely well in a driverless-car economy, if plenty of short-wait hailable options exist at all stations. Given driverless cars won't balk at short-distance drives (like taxi drivers driving you 6 blocks to a subway station, instead of directly to destination), they would be clockwork reliable no matter how short your journey was. Some self driving car services will even be government operated services. The pundits that say rapid transit will disappear in a driverless economy, is just blowing hot air.

Sheer numbers and road capacity pretty much guarantees continued need for rapid transit options. It definitely will impact transit, it will strengthen the raison d'etre for rapid transit along central corridors, it will restructure inefficient transit (e.g. circuitous suburban bus routes with few passengers), but it isn't going to kill transit as a whole.
 
Last edited:
I would think that the utopian vision is far more likely. If an autonomous car can arrive quickly at your house, and be cheaper than a cab (since there is no driver), why bother owning? I could see condos having fleets of self-driving cars parked in the condo garage, making getting a car as fast as if you owned it. Sure, there would still be folks who want to own their own car, but I think we'd see a societal shift very quickly to transportation as more of a service one rents, and less of a product one owns. And since autonomous vehicles are more efficient when there aren't other unpredictable humans driving on the roads, I can easily see a time when only certain roads are designated for human drivers, and all others only legally allow self-driving vehicles, further decreasing the attractiveness of non-autonomous vehicle ownership.'

I know that some people feel a romance about cars, just as some people feel romance about horses. But as horses have gone from ubiquitous to being a luxury that only few can afford, I think the same will be true of cars. And I think this shift will happen very quickly once the technology is actually good enough.
 
The issue of bad weather is obviously especially pertinent in our climate, and I think it is no accident (ha!) that almost all the autonomous vehicle work is being done in sunny California.

California (and Boston) is where the appropriate tech employees were but many companies (including Google) are shifting testing work to the UK due to significantly fewer regulation hurdles.

I don't doubt they'll continue testing in California too but they're going to be getting some good poor weather testing too.
 
  • Moral -- Cars that are faced with an unavoidable fatality decision are going to decide whether to save the occupants or pedestrians. Picture the scenario of a baby stroller suddenly running in front of the car at the last second, from behind roadside newspaper boxes (unavoidably unseen by the car's sensors until too late; now a fatality has to happen). Car must instantly decide to crash into baby stroller OR suddenly veer into a parked car/lamppost 1 meter to the side. Legally solve this. Now consider the sole occupant of car is your child being soccermomed unaccompanied to school. Whose life goes? Whose Responsibility? Legislative? Insurance? Etc

It's a though experiment which in practice doesn't seem very relevant. Autonomous vehicles, like humans, won't ever be making their decisions based on full information. They'll never be able to know everything about their surroundings. Maybe that women's stroller is just being used as a shopping cart and there is no baby in it, meanwhile maybe there are two kids in the parked car.

If an autonomous vehicle detects an obstruction that will likely cause an unavoidable accident, the best solution will always be to just apply the brakes instantly. Even if a collision is unavoidable reducing speed from ~40km to ~20km and informing trailing vehicles to brake immediately to avoid further collisions would be the morally preferable course of action.

Expecting cars to perform some moral-kamikaze is more likely just to cause automated versions of that accident were a woman caused a pileup trying to avoid hitting some ducks.
 
Last edited:
It's a though experiment which in practice doesn't seem very relevant. Autonomous vehicles, like humans, won't ever be making their decisions based on full information. They'll never be able to know everything about their surroundings. Maybe that women's stroller is just being used as a shopping cart and there is no baby in it, meanwhile maybe there are two kids in the parked car.

If an autonomous vehicle detects an obstruction that will likely cause an unavoidable accident, the best solution will always be to just apply the brakes instantly. Even if a collision is unavoidable reducing speed from ~40km to ~20km and informing trailing vehicles to brake immediately to avoid further collisions would be the morally preferable course of action.

Expecting cars to perform some moral-kamikaze is more likely just to cause automated versions of that accident were a woman caused a pileup trying to avoid hitting some ducks.

Your example is the easy one for the computer. But I read a better example in an article a month or so ago. Say a car is on a mountain road and all of a sudden a person is blocking the road. There are 2 options (and only 2). One is to hit the breaks but still hit the pedestrian and there is a 100% chance they will be killed. The other option is to steer off the cliff and kill the people in the vehicle. What would the computer be programmed to do?

Does this change if the computer does not know if there are 1 or 5 people in the car?
Does this change if the pedestrian is breaking the law and jumped on the road?

It does create a moral question that now has to be decided by society vs the person sitting in the car.
 
Your example is the easy one for the computer. But I read a better example in an article a month or so ago. Say a car is on a mountain road and all of a sudden a person is blocking the road. There are 2 options (and only 2). One is to hit the breaks but still hit the pedestrian and there is a 100% chance they will be killed. The other option is to steer off the cliff and kill the people in the vehicle. What would the computer be programmed to do?

The car wouldn't be able to make a calculation like "100% chance they will be killed." That's unknowable. Nor would the car be able to tell if fewer people would be killed by driving off the road; imagine a school at the base of the hill.

These question's don't accurately reflect the real world. Drivers aren't expected to grapple with the trolley problem because it doesn't actually happen. Even if it did, the only way to know would be retrospectively.

In reality, a car would be expected to take reasonable precautions to avoid striking the man blocking the road. Even reducing impact speed from 100 kph to 65 kph would reduce the average risk of the pedestrian's fatality from 100% to 50%. That's a moral and utilitarian jackpot that would always outweigh some kind of outlandish attempt to avoid collision by executing James Bond maneuvers.
 
Last edited:
How will it affect public transit? Rapid transit corridors will probably boom (e.g. subways, commuter trains, high-traffic bus/LRTs, high speed trains) with self-driving minibuses/Ubers/transit/taxis/etc becoming one of the many last-mile connecting options, while low-traffic suburban buses may die (e.g. UberPOOL style services may be more profitable for lower fares per person, and TTC/GO may run UberPOOL style dial-a-bus services much more efficiently than a fixed-route suburban bus, for example -- assuming they keep dynamic pickup/dropoff routes as ultraefficient as possible, and number of occupants per vehicle low -- then it can probably beat the transit economics and transit performance performance of a low-usage suburban bus route).
.....
Sheer numbers and road capacity pretty much guarantees continued need for rapid transit options. It definitely will impact transit, it will strengthen the raison d'etre for rapid transit along central corridors, it will restructure inefficient transit (e.g. circuitous suburban bus routes with few passengers), but it isn't going to kill transit as a whole.

The beautiful part about automation is that road capacity increases. Cars will be able to zip along the Gardiner closer together, at higher speeds and with more lanes than before. The Gardiner could become a 8 lane road without changing the width with no centre median. That means the number of lanes in each direction can automatically shift based on the feeds from all the cars!

http://www.tampa-xway.com/Portals/0/documents/Projects/AV/TAVI_8-CapacityPinjari.pdf

It is estimated that a single lane of traffic on a highway will have an increased throughput from 2200 cars to 4000 cars per hour. And if you can squeeze more lanes in you can increase it even further. And individual mistakes cause >90% of all accidents which mostly will be eliminated which will reduce further traffic issues.

All this means we can have more people driving to work even in rush hour. And you can know exactly how many cars are driving at the same time. So in the morning you can say I want to leave the house at 8:00am and the car will tell you if there is capacity on the road system...if there isn't it will just not let you on the road (and the capacity can be managed with variable tolling).

Where there are significant capacity constraints (likely where high order transit exists now) there will still be a need for higher-order transit or car pooling. Those lines that are really slow or were showpieces for a politician will have very significant declines in use.

I totally agree with you on bus systems. The feeder systems will be taken over by either a transit pooling service or a dedicated taxi pool (whichever you can afford). Great for the 'burbs feeding into the Go Train system!
 
That's a moral and utilitarian jackpot that would always outweigh some kind of outlandish attempt to avoid collision by executing James Bond maneuvers.
The default of "only always try to brake in front of imminent collision" isn't going to clear-cut it in 100% of all situations.

There comes an overwhelming moral scenario such as when a school bus (50 students) suddenly enters the path of a self-driving semi truck (1 truck driver sleeping in cabin). At some point, it becomes overwhelmingly more favourable to attempt to avoid the school bus, and crash into a rock median or rural forest that obviously has no humans in them.

There comes a point where it's so clear-cut you must need to attempt to avoid, when mathematically you can only decelerate to, say 50kph from 100kph given the time constraint of sudden discovery the bus veers into your lane -- i.e. a tire blowout -- in an extenuating situation (e.g. school bus suddenly crosses median at last fraction of second, because of tire blowout).

Imagine a scenario where we discover a self-driving truck killed 15 children because it made a flawed moral decision, and we start to legally wrangle over it -- delaying the permission of full Level 4 freedom on all roads Canada-wide...

Sometimes the moral is so clear cut, that you must avoid just like a human would try to. It's not James Bond stuff, buddy.

This is the kind of stuff that governments and legal will wrangle over for a while, before permitting full Level 4 freedom for all vehicles of all sizes (possibly circa 2075). The transition is going to be gradual and long term, very cautious, so many of us may not get the Level 4 freedom in some parts of the country within our lifetime -- depending on progress -- while other parts will have it self-valeting empty vehicles, as quickly and early as possible (possibly circa 2035?). And Level 3 (human-supervised self-driving cars, that allows you to text or watch movies) well before then, possibly as soon as this decade. But Level 4 is a whole different ballgame -- the ability to safely taxi children and drunks (who may vandalize their ride) unattended.

The moral issues are going to be fought over, challenged, studied, researched, re-studied, and re-fought over, before Level 4 wins freedom on 100% of roads Canada-wide (all the way to self-driving buses, transit, taxis, riding on public roads, including downtown cores, school zones, in middle of blizzards, etc).
 
Last edited:

Back
Top