Two Spoke Forums banner

1 - 11 of 11 Posts

·
Deranged Touring Cyclist
Joined
·
5,208 Posts
Discussion Starter · #1 ·
It appears the first fatal accident involving a pedestrian and an autonomous automobile has happened in Tempe, AZ. While I'm interested in developments surrounding autonomous vehicles in general, this one caught my attention a bit more due to the fact that although the victim is reported to have been hit while crossing a street on foot and outside a crosswalk, she was hit while walking her bike.

I haven't seen a lot of information on how autonomous vehicles identify and avoid bicyclists and pedestrians, but as a cyclist who regularly plays in the streets with cars, I'm really interested in the same.

There's been a great deal of press surrounding the idea that autonomous automobiles are more safe than those operated by humans. I am open to the idea that this is correct, but uncertain that the evidence currently supports such a conclusion.

What do you think? I imagine most every road going cyclist has dealt with inattentive and distracted drivers. I like the idea of electronics ceaselessly scanning the road ahead, seeing my bicycling @$$ in the lane ahead, and steering to avoid me.

I fear that in the short term, at least, we cyclists may be looking at struggling to be 'seen' by a range of sensors which may or may not bear any relation to human eyes. How shall these sensors recognize us? I ride with hi-vis and a range of lights to attract the attention of human drivers. How to attract the 'attention' of autonomous systems created and programmed by a variety of independent companies?

Things may get interesting in times to come.
 

·
Two skinny J's
Joined
·
21,353 Posts
It was bound to happen and I figured when it did everybody would have a general negative reaction to it.

I'm personally not a big fan of all the automation, and autonomy.

It just feels like this, We are the Borg. You will assimilate.

 
  • Like
Reactions: fpl1

·
Registered
Joined
·
23 Posts
It was bound to happen and I figured when it did everybody would have a general negative reaction to it.

I'm personally not a big fan of all the automation, and autonomy.

It just feels like this, We are the Borg. You will assimilate.

At least the Borg had a hive mentality. My fear is that we're joining Sarah Connor on the road to Terminator.
 

·
Administrator
Joined
·
9,948 Posts
Personally, I don’t think they’ll catch on in this country. While it’s neat to have your car drive you around while you just sit there, our society is in too much of a hurry to get nowhere. These vehicles will drive the speed limit and stop at all lights and stop signs unless manually overridden and that defeats their purpose. If the operator gets complacent and stops paying attention to what’s going on and has to stop the car in an emergency, his/her reaction time will drastically increase and the crash will, most likely, still happen.
 

·
Registered
Joined
·
77 Posts
This technology, while having been in the news more than a decade ago, is still relatively in its infancy. Consider it took almost 100 years to get electronic fuel injection right, and petroleum is still much to valuable to be burning. Mistakes and glitches are inevitable.

It's also important to understand that very much like an ABS pump I fixed after it failed and caused an accident, no matter how long the technology matures, it will never be perfect, and it will potentially cost lives. The question is if it will save far more lives and prevent far more property damage than it takes and causes. I think in time it will, but I have seen nothing to make me confident we are there yet.
 

·
Premium Member
Joined
·
7,064 Posts
^ I agree with you. It will take some time but generally I believe we will continue to see advances in this technology which will improve safety. It might be a bumpy road getting there but I suspect companies will continue to make progress on the effectiveness and safety of this tech.
 

·
Old, fat, and slow
Joined
·
789 Posts
Eventually almost all urban/suburban vehicles will be self-driving. The biggest reasons are traffic and safety. Traffic jams will be minimized because things like merging and letting vehicle make turns won’t be impossible, as they seem to be with ignorant human drivers. Even if the cars travel the speed limit, trips could be shorter because of increased efficiency.

Also, computers tend not to get drunk, do drugs, turn around to talk to people in the back seat … they text while driving but have that capacity. They don’t carry firearms and shoot people who passed them.

Uber has and has had the worst safety record---unable to cover 13 miles without needing the driver to intervene, while Waymo managed a Slightly greater 5200 miles without an incident.

Also, the Uber sensors were either not turned on were not connected to the brain. Lidar could have seen that woman walking from a huge distance, dark or not, but if it did, the computer never registered or reacted.

This was a tragedy and also a failure of a really shoddy AV system. Meanwhile, Waymo has been operating taxis with No human drivers safely for over a month and there has been no news because Waymo spent the money building a good system and did the miles testing it before they turned it loose.

And no, I don't own stock in Waymo, it is just that I have been researching this a little since it happened.
 

·
Two skinny J's
Joined
·
21,353 Posts
Eventually almost all urban/suburban vehicles will be self-driving. The biggest reasons are traffic and safety. Traffic jams will be minimized because things like merging and letting vehicle make turns won’t be impossible, as they seem to be with ignorant human drivers. Even if the cars travel the speed limit, trips could be shorter because of increased efficiency.

Also, computers tend not to get drunk, do drugs, turn around to talk to people in the back seat … they text while driving but have that capacity. They don’t carry firearms and shoot people who passed them.

Uber has and has had the worst safety record---unable to cover 13 miles without needing the driver to intervene, while Waymo managed a Slightly greater 5200 miles without an incident.

Also, the Uber sensors were either not turned on were not connected to the brain. Lidar could have seen that woman walking from a huge distance, dark or not, but if it did, the computer never registered or reacted.

This was a tragedy and also a failure of a really shoddy AV system. Meanwhile, Waymo has been operating taxis with No human drivers safely for over a month and there has been no news because Waymo spent the money building a good system and did the miles testing it before they turned it loose.

And no, I don't own stock in Waymo, it is just that I have been researching this a little since it happened.
Interesting, never heard of Wayno. Were I currently live we now officially have one Uber :D Like it that way. We use Uber frequently up in DC and my sons uses it often in Richmond.

Good point about the "zipper" effect the cars will have in traffic tho, They'll get hacked someday soon I will bet :D
 
  • Like
Reactions: MilesR

·
Deranged Touring Cyclist
Joined
·
5,208 Posts
Discussion Starter · #10 ·
Interesting, never heard of Wayno. Were I currently live we now officially have one Uber :D Like it that way. We use Uber frequently up in DC and my sons uses it often in Richmond.

Good point about the "zipper" effect the cars will have in traffic tho, They'll get hacked someday soon I will bet :D
Waymo is Google's effort at developing self-driving cars. Interestingly, there was just a whole intellectual property legal thing between the two companies, which Uber paid to settle (lost).

While competition is generally to be encouraged, I wonder how multiple independently developed autonomous vehicle types will interact, both with one another and with non-autonomous vehicles such as holdout auto drivers, bicyclists, pedestrians, etc.

I also wonder how malevolent humans will choose to interfere: if people are willing to shine laser pointers at aircraft despite the obvious danger of causing a crash by blinding the pilot, how many will seek out ways to interfere with or defeat autonomous automobile sensors? That there will be such ways is virtually guaranteed by the combination of new technology, sensors, and evolving software.

There are also software bugs and unanticipated consequences of programming and hardware interactions. Either of those last may have been responsible for the OP's tragic accident. It seems clear that the vehicle's systems utterly failed to detect the woman and her bike, given the reported lack of braking prior to the crash.

I'm all for added safety, but I've spent too much time professionally unf%#king software and hardware problems in the context of office environments to think that multiple independently developed software/hardware packages will simply work on their own in the real world, much less with one another. Over time, sure. In the short term? Not so much.

These are interesting times in which we live.
 

·
Registered
Joined
·
77 Posts
They still can't figure out how to make a car that won't convince itself to run poorly if a sub system stops sending the signals the computer expects, or the computer itself gets faulty which is extremely common on general motors vehicles. These things are readily comparable to the computer being drunk or having dementia.

Corrosion, pack rats, plastic degradation, poor designs, extreme weather, these things can be counted on to disrupt the proper function of these driverless vehicles, and as a best case scenario, they will in the decades to come make them fairly safe but probably less reliable than current vehicles because a failed pre drive self check will mean the computer won't let itself run.

It will be a whole new set of headaches with plenty of the same old ones.
 
1 - 11 of 11 Posts
Top