Bike Forums

Bike Forums (https://www.bikeforums.net/forum.php)
-   Advocacy & Safety (https://www.bikeforums.net/forumdisplay.php?f=8)
-   -   The looming threat of autonomous vehicles (https://www.bikeforums.net/showthread.php?t=1174622)

holytrousers 06-02-19 05:15 PM

The looming threat of autonomous vehicles
 
when autonomous vehicles become the norm, cyclists will be the first to experience the dramatical realization that our civilzation took the wrong turn.
i have been thinking about this inevitable fate with great concern, and i thought i would ask you what are your thoughts about it ?
by being the most vulnerable users on the road, how do you think the dominance of these autonomous machines will impact our psychology and everyday life ?

ridelikeaturtle 06-02-19 05:39 PM

It would be my hope that the autonomous vehicle would place environmental safety higher in its "hierarchy of needs" than determining whether or not they arrive "on time". I could be wrong on that, especially for delivery vehicles... which is where I'd hope good legislation and regulation comes into force.

If not, then it'll take a few fatal accidents to get people fired up about it.

However, I'd expect autonomous vehicles to be a saviour for road safety. Algorithms don't rust; programming doesn't fall asleep, and (with the above caveat) aren't going to pass a cyclist while approaching a blind turn or crest of a hill because they're impatient.

That'll make cycling safer, and that's gotta be a huge incentive to cycle.

350htrr 06-02-19 07:14 PM


Originally Posted by holytrousers (Post 20958916)
when autonomous vehicles become the norm, cyclists will be the first to experience the dramatical realization that our civilzation took the wrong turn.
i have been thinking about this inevitable fate with great concern, and i thought i would ask you what are your thoughts about it ?
by being the most vulnerable users on the road, how do you think the dominance of these autonomous machines will impact our psychology and everyday life ?

I would say you would be 100% wrong... In the long run... Certainly for the first while there would be a higher % of cyclist's/pedestrians that would end up being run over, and hurt or killed, even other motorists... BUT, when it get all sorted out it will be a whole lot safer for everyone... JMO...

JW Fas 06-02-19 07:21 PM


Originally Posted by holytrousers (Post 20958916)
when autonomous vehicles become the norm, cyclists will be the first to experience the dramatical realization that our civilzation took the wrong turn.

How did you arrive at this?

gregf83 06-02-19 07:23 PM

With 36,000 automotive deaths/yr in the US alone I think the bar is awfully low for autonomous vehicles to clear. On average, they're probably safer now and will only get better, unlike humans who continue to kill cyclists and pedestrians at an alarming rate.

tyrion 06-02-19 07:35 PM

I'll feel much safer when we get human driven vehicles off the road.

Shimagnolo 06-02-19 08:09 PM

I could argue either side of this debate.

On one hand, it seems like an autonomous system should be safer since:
- It won't get road rage.
- It won't fall asleep.
- It won't drive drunk
- It won't focus its attention on a cell phone instead of driving.

On the other hand we have seen:

- An Uber self driving car that detected a pedestrian, but had auto-braking disabled, and had alerts to the safety driver disabled.

- A Boeing aircraft designed to over-ride the pilots when a single angle-of-attack sensor fails, even with autopilot off, and this was kept secret from the pilots.

- An auto mfgr (Tesla), which claims on its website that its autos have "Full Self-Driving Capability", when they do not: https://www.tesla.com/modelx/design#autopilot


So as long as we have fools in corporate mgmt, all bets are off.

Rick 06-02-19 08:52 PM

It is the same humans that run us over that are programing those autonomous vehicles and deciding upon what is cost effective on its ability to detect problems. Us being the problem.

rossiny 06-02-19 09:01 PM

I would feel much safer if we had public transit , high speed rail and trains Linked to small buses running every 10 minutes or so and completely eliminate cars. Cars will eventually destroy the earth and man kind any way. In my opinion. Combined with all the rest of the manufacture and pollution. I just cant beleive it would even be considered. I mean just look at how glitchy your computer is a times. Or how glitchy your phone internet service is, etc, etc. The only problem with cars and biking is that they are not separated, and the distracted driver, with looking at his phone or texting. I really never thought talking on phone was a danger, especially with a flip phone. You keep in the same spot, you flip open and talk. You can still have your head up and see all the road. Now dialing, different story there. .
I guess a cement barrier like the one dividing the highways would be awesome , but that is not gonna happen

tyrion 06-02-19 09:01 PM


Originally Posted by Rick (Post 20959214)
It is the same humans that run us over that are programing those autonomous vehicles and deciding upon what is cost effective on its ability to detect problems. Us being the problem.

Humans do manual accounting and humans make accounting software. The human made accounting software is much (much) less error prone than human accountants.

holytrousers 06-02-19 09:46 PM

I think we can agree on two obvious axioms related to this question :
First one is that cars are machines that, besides transporting people, represent a real threat to other road users.
The moment we delegate the responsibility of controlling a killing machine to an obscure algorithm totally lacking the ability to empathize with a human being or understand and predict his behaviour, we are de facto deliberately submitting ourselves to an absolute and anonymous control that's beyond the influence and understanding of normal humans.
The second point is that currently danger on our roads stems mainly from human irresponsibility. The real issue here is how to educate our children and adults to become responsible citizens. Substituting a problem with another problem is certainly no the simplest way to proceed.
I always feel strongly skeptical when i have to deal with a "safety" argument : it has been misused too often to let our fears hinder our rational thinking on this matter. We all remember how mass surveillance programs and military invasions have been enforced for security concerns.
Let's teach children responsibility instead of filling their heads with useless dates and obsolete formulae.. Replacing people with robots has nothing to do with improving safety on our roads, that's just marketing and manipulating the public opinion.

jon c. 06-02-19 10:00 PM

I don't feel threatened by the prospect. It will be quite a while before the tech is widespread and it will be safer than humans are now.

noimagination 06-03-19 06:20 AM


Originally Posted by holytrousers (Post 20959256)
I think we can agree on two obvious axioms related to this question :

First one is that cars are machines that, besides transporting people, represent a real threat to other road users.


OK, postulated.




Originally Posted by holytrousers (Post 20959256)
The moment we delegate the responsibility of controlling a killing machine to an obscure algorithm totally lacking the ability to empathize with a human being or understand and predict his behaviour, we are de facto deliberately submitting ourselves to an absolute and anonymous control that's beyond the influence and understanding of normal humans.


I fail to see how this follows from axiom #1 . The sentences seem totally unrelated. Also, you use many words loaded with emotional connotations, which significantly undermines your pose of logical argument.




Originally Posted by holytrousers (Post 20959256)
The second point is that currently danger on our roads stems mainly from human irresponsibility. The real issue here is how to educate our children and adults to become responsible citizens. Substituting a problem with another problem is certainly no the simplest way to proceed.

I always feel strongly skeptical when i have to deal with a "safety" argument : it has been misused too often to let our fears hinder our rational thinking on this matter. We all remember how mass surveillance programs and military invasions have been enforced for security concerns.

Let's teach children responsibility instead of filling their heads with useless dates and obsolete formulae.. Replacing people with robots has nothing to do with improving safety on our roads, that's just marketing and manipulating the public opinion.


Agreed. If you have some practical process for eliminating human irresponsibility, or even of significantly reducing it (and what is more, for getting such a process implemented), I'm sure everyone would love to learn about it.


The point is, humans are fallible. I understand that it frightens people when they perceive that they have given up control, even if that sense of control is illusory. It is certainly possible that the people designing the programs for self-driving cars could make mistakes that could cause fatalities. It is certainly possible that there could be random faults in the software or hardware occurring after installation and testing that could cause fatalities. It is certainly possible that users could misuse the autonomous drive feature. It is certainly possible that malicious hackers could introduce faults that could cause fatalities. Would the cumulative risk posed by autonomous vehicles be greater than the risk posed by millions of vehicles operated by millions of humans all fallible in different ways at different times? That remains to be seen.

Rick 06-03-19 06:41 AM


Humans do manual accounting and humans make accounting software. The human made accounting software is much (much) less error prone than human accountants.
Humans also make decisions based on cost effectiveness. They will make there automated systems as cheap as possible. Tesla and other companies have already started this by allowing systems that are inadequate to the task. Life is cheap to these people. Mercedes and other car manufactures are already negotiating with the insurance companies on not being held responsible in any way if there machines nail a pedestrian or bicyclist. I personally do not want to be one of there experiments in cost effectiveness.

mixteup 06-03-19 08:48 AM

Although in theory, they could solve some problems, ever since autonomous vehicles came on the radar big time, my concern is how easily were their systems able to be hacked.
I believe it's already proven that they can be to some extent, through the gps and other sensors they use.
So it seems inevitable that it will be done, either for "fun" or deliberate malice.
Could be a real problem, once a lot of them are on the road - Hopefully, I'm wrong

tyrion 06-03-19 10:29 AM


Originally Posted by Rick (Post 20959525)
Humans also make decisions based on cost effectiveness. They will make there automated systems as cheap as possible.

Incorrect. A for profit company tends to maximize profits. Making safety critical systems as cheap as possible isn't the way to do that, and most people in the business know that. A parachute company that skimps on safety to make the cheapest parachute doesn't stay in business long.

In the automobile business, safety is good business.

bobwysiwyg 06-03-19 10:36 AM


Originally Posted by mixteup (Post 20959712)
Although in theory, they could solve some problems, ever since autonomous vehicles came on the radar big time, my concern is how easily were their systems able to be hacked.
I believe it's already proven that they can be to some extent, through the gps and other sensors they use.
So it seems inevitable that it will be done, either for "fun" or deliberate malice.
Could be a real problem, once a lot of them are on the road - Hopefully, I'm wrong

My thoughts on this are, good luck. If major banks, corporations and NSA can't guard against hacks and intrusions, what chance do the auto makers have with their "autonomous" vehicles?

Leisesturm 06-03-19 11:15 AM

No one yet has mentioned the socio-economic impact of 3 million unemployed over the road truckers being dumped onto the society if and when AV's transform that industry! Add to that taxi and limousine drivers, local delivery drivers. THAT is the reason why fully autonomous vehicles are the "wrong turn" for civilization. Human programmers will NOT program AV's to hit cyclists! The poster that pointed that out is about the only person with their head on straight this morning.

The lane monitoring, collision detection and automatic braking systems in a lot of production cars should be all the automation allowed in motor vehicles until true Artificial Intelligence is achieved. The ideal platform for a driver is a vehicle that will make sure s/he is awake and alert (online), maintaining their position in their lane (online), not about to hit something (online) and, if in an out of control situation not overcorrecting or overbraking (online). At present these systems are not penetrating down to entry level car models which are the majority of cars sold. That should change. True AI is decades away, if ever. We should hope for, if ever. Guaranteed, it is more likely that the human programmers of AI will attempt to use it to control or kill humans that their governments don't like. This is far more likely than that cyclists will be in any way targeted. AV's will make for more predictable roads and it MAY make them safer to cycle on but it MAY be that in order for that to happen that the roads will need to be cleared of vehicles that cannot be controlled via AI. Be careful what you wish for.

AV's are being developed to replace human drivers. This is for economic NOT social reasons. We could do a lot by making existing drivers safer but they would still need to be paid if they are driving for a living. So AV's are being rushed to market in a very primitive state because some people can't wait for them to reach a proper level of development before being turned loose on civilization.

tyrion 06-03-19 11:52 AM


Originally Posted by Leisesturm (Post 20959948)
No one yet has mentioned the socio-economic impact of 3 million unemployed over the road truckers being dumped onto the society if and when AV's transform that industry! Add to that taxi and limousine drivers, local delivery drivers.

That's going to be a massive problem. The disruption is going to be big and painful. But there's no avoiding it - best course of action is to plan well ahead to soften the blow. Do not sign up for an expensive truck driving course.

squirtdad 06-03-19 01:42 PM

IMHO...... it will be a long time before self driving vehicles are wide spread and even longer before there are self driving vehicles without steering wheels and a driver take over options

Practical (it's the little things will get you) issues like
  • keeping the sensors clean (has been a big issue, will continue to be an issues)
  • dealing with bad weather.....like heavy rain, snow, ice, hail you have not seen a lot of testing out side of sunbelt states
  • can't always depend on satellite and cell connectivity. I get this just driving from San Jose to Eureka in Cal.
  • dependence on clear and consistent pavement markings. this will be cost issue for rural areas....and people always seem to forget about rural areas (i.e places like Montana whey you drive 200 miles or more one way to play highschool sports

Functional issues

IMHO it is not higher speed freeway driving that is the achilles heel for self driving, it will be the slow speed complex everyday situations the will be hard as an example my commute on my bike (or car) during morning school commute hours with with 5 schools, crazy parents running late, kids walking, people walking dogs, kids bicycling, teenagers talking on their phone and otherwise oblivious. Narrow streets where often one car has to wait for another, limited visibility, 1 stop sign, 1 4 way stop sign, 3 traffic lights, 2 crosswalks across 4 lane street. people hanging a right then crossing over 2 lanes in a 100 feet, multple right turns across a bicyclist. I fail to see any current self driving vehicle being able to handle this. and when I talk to people who work in the field....they say...oh take a different route, but the problem is to for parents to get to schools they have to go this way

legal issues:
  • the law is not ready for self driving vehicles.
  • who gets a ticket if a self driving vehicle has a traffic violation the owner of the vehicle, the occupnat, the company that made the vehicle, the company that made the software and /or hardware?
  • which leads to who get's sued if there is an injury

Road Fan 06-03-19 01:50 PM


Originally Posted by Rick (Post 20959214)
It is the same humans that run us over that are programing those autonomous vehicles and deciding upon what is cost effective on its ability to detect problems. Us being the problem.

I can't agree, because I am a person involved in the design of autonomous vehicles and helping to decide how decisions are made, and I have not run over any cyclists ever in my life.

It's possible to paint a wall with too broad a brush.

Rick 06-03-19 01:54 PM


Incorrect. A for profit company tends to maximize profits. Making safety critical systems as cheap as possible isn't the way to do that, and most people in the business know that. A parachute company that skimps on safety to make the cheapest parachute doesn't stay in business long.

In the automobile business, safety is good business.
This is true in many industries. We are talking about stupid here. Uber getting off of the hook for murder in the road with sub standard equipment and backup driver. Tesla with the sensors that couldn't recognize a large truck and killing the driver watching Looney Tunes. In a perfect world you would be correct. This is no ware near a perfect world. The corporations want to make money and they are not going through the effort of making sure there equipment is good enough.

Road Fan 06-03-19 02:06 PM


Originally Posted by holytrousers (Post 20959256)
I think we can agree on two obvious axioms related to this question :
First one is that cars are machines that, besides transporting people, represent a real threat to other road users.
The moment we delegate the responsibility of controlling a killing machine to an obscure algorithm totally lacking the ability to empathize with a human being or understand and predict his behaviour, we are de facto deliberately submitting ourselves to an absolute and anonymous control that's beyond the influence and understanding of normal humans.
The second point is that currently danger on our roads stems mainly from human irresponsibility. The real issue here is how to educate our children and adults to become responsible citizens. Substituting a problem with another problem is certainly no the simplest way to proceed.
I always feel strongly skeptical when i have to deal with a "safety" argument : it has been misused too often to let our fears hinder our rational thinking on this matter. We all remember how mass surveillance programs and military invasions have been enforced for security concerns.
Let's teach children responsibility instead of filling their heads with useless dates and obsolete formulae.. Replacing people with robots has nothing to do with improving safety on our roads, that's just marketing and manipulating the public opinion.

Safety arguments? Useless dates? obsolete formulae? So no crafts of argumentation or logic, no history (it's rather more than dates, doncha know?), and no quantitative physical science? Perhaps you want education to be limited to the crafts and manual farming? Quilting, sewing, cooking, plowing, carpentry, etc?When you no longer have computers that function, because the next generations can't maintain any equipment more sophisticated than the bellows needed by the local blacksmith, how will you type on the Internet?

Road Fan 06-03-19 02:30 PM


Originally Posted by tyrion (Post 20959864)
Incorrect. A for profit company tends to maximize profits. Making safety critical systems as cheap as possible isn't the way to do that, and most people in the business know that. A parachute company that skimps on safety to make the cheapest parachute doesn't stay in business long.

In the automobile business, safety is good business.

Totally agree! Actual safety is also needed to minimize actual crashes, which minimizes not only loss of life and health, but corporate and societal cost in the form of class-action suits. The threat and unpredictability of such expenses tends to frighten companies into putting in place processes for engineering safety into products, and to develop standards for how to perform and check the execution of those processes. From my viewpoint, there are really hard problems for individual vehicles, like detecting, tracking, and predicting humans in the vehicle, pedestrians and cyclists (just to name a few that are close to our hearts here). But with AV's we have to think of what they might do in traffic or crazy things owners might ask them to do, such as to drive around the block all day until the owner leaves work and wants to have the car pick him up to drive home. What if all 50 or 500 workers in the office try to do this? Solution to this is to greatly expand public transit so workers have a more efficient way to get home. But who pays for that?

And on and on and on ... We cyclists are a problem but not the only one. And safety on a cycle is not the only system fail that could ruin your day.

berner 06-03-19 03:09 PM

I've just finished reading a biography of Elon Musk which to a large extent is also a biography of Tesla. There were two road fatal incidents with Tesla vehicles where the vehicle failed to recognize trucks crossing in front of the oncoming Tesla which then drove under the trucks shearing off the upper section of the cars. As explained in the book, doppler radar was used to identify movement in front of the Tesla. Evidently, doppler shifts are not good at detecting movement across the direction of travel so the trucks were never "seen" by the radar. The book did not say whether this system has been corrected. What was clear is that billions of dollars are at stake and Tesla is running out of time and money, as usual.


All times are GMT -6. The time now is 08:06 AM.


Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.