But with regard to police pulling a sign out of their trunk which has been sitting there for a week... they can’t. I’ve explicitly stated that any QR code means absolutely nothing by itself. So you have a copy of a meaningless sign? Tell me what the risk is?

Then were is this sign coming from that you stated:

First responders needn’t generate a new QR code nor anything else you suggest. They pull a sign out of the trunk of their cruiser and put it up. At a minimum it instructs the auto car to go to manual mode. Otherwise, a dispatcher can update specific instructions.

You can't have it both ways -- they either just pull a sign out of their trunk or they don't. Which is it?

If you are saying that the sign has to be activated each time it is used, then that runs into two of the issues I've already touched on -- it requires the first responders to spend time doing administrative tasks instead of attending to the incident and it requires that all incidents occur in places that have sufficient connectivity for the necessary communications to occur.

Really, these similar issues always appear after the fact. But if the designers looked ahead, it’s a lot easier to design away a problem than try to retrofit a solution.

This is actually an interesting question that I had a long and fascinating discussion with our recent NSA Visiting Professor. I mentioned that it seemed to me that all of these new technologies always bolt on security as an afterthought and never seriously design it in from the beginning. He responded that bolting it on after the fact is the way it should be (or perhaps a more accurate description is that it's the way it has to be). That set me back on my heels (given his background) but over the course of the discussion he pretty much swayed me to his side -- or at least convinced me that he had a very valid view. He pointed out system after system that has attempted to build security into it from the beginning and noted that none of them have ever seen the light of day, including a secure operating system that the NSA has been trying to develop for a couple of decades now. Systems that do see the light of day usually have a relatively basic security foundation and then, as the issues are uncovered, they are addressed through patches and upgrades.

The discussion started over IPv6 and how originally the utopian-dreamers that developed it assumed that you could just assign each person a huge block of addresses or assign a different block of IP addresses to each square millimeter of the Earth's surface (or both) and how later others pointed out the privacy and tracking issues with either approach. The solutions are being developed as after-the-fact bolt-ons. When I claimed that it would have been nice had the original developers at least given some thought to issues such as this, he commented that if they had then they would probably still be going back and forth about how to address them and we would still be years away from seeing IPv6 implemented at all (if ever).

 

Interesting observation from your NSA friend. I will chew on that for a while. My gut reaction is that a hybrid approach is necessary. Don’t build in security; but identify where security is necessary during design. It’s kind of like this old math joke.



As far as pulling a sign out of the trunk of s police cruiser, I believe I addressed that issue in my post. Of course we don’t want to waste first responders time with administrative crap. But how did they get there in the first place? Usually there are dispatchers, whose job is administrative in the first place. No administrative tasks required of first responders.

And finally, with respect to your irrelevant concern that “all incidents occur in places that have sufficient connectivity for the necessary communications to occur”, I have two replies. 1) if an autonomous vehicle is operating, then by definition sufficient communications exist. 2) the sign doesn’t need to communicate anywhere other than to the autonomous vehicle. It’s a read-only device. By itself, it only contains one piece of information and by itself does not require communication.

 

Autonomous means self-governing. While external communications is nice to expand the range of sensors, the requirement of totally local control based totally on local sensor inputs is necessary to differentiate it from automated/cooperative control using a system of external communications links (or a driver) with latency issues that could delay the needed response time to avoid collisions.
https://web.archive.org/web/2017051...ke.sk/hudecm/PDF_PAPERS/Intro-Aut-Control.pdf
https://www.euroncap.com/en/vehicle-safety/the-rewards-explained/autonomous-emergency-braking/

 

Interesting observation from your NSA friend. I will chew on that for a while. My gut reaction is that a hybrid approach is necessary. Don’t build in security; but identify where security is necessary during design.

That's the general position I've come to -- that there needs to be a proper balance (and what's proper is going to be very application-specific). But I do see the very real potential for problems stemming from the increasing security-consciousness of the customers (not to mention government regulators and the legal system). I can easily image getting to the point where security requirements placed in the initial requirements will bring innovation and implementation to its knees. Unintended consequences.

As far as pulling a sign out of the trunk of s police cruiser, I believe I addressed that issue in my post. Of course we don’t want to waste first responders time with administrative crap. But how did they get there in the first place? Usually there are dispatchers, whose job is administrative in the first place. No administrative tasks required of first responders.

And finally, with respect to your irrelevant concern that “all incidents occur in places that have sufficient connectivity for the necessary communications to occur”, I have two replies. 1) if an autonomous vehicle is operating, then by definition sufficient communications exist. 2) the sign doesn’t need to communicate anywhere other than to the autonomous vehicle. It’s a read-only device. By itself, it only contains one piece of information and by itself does not require communication.

A properly autonomous vehicle requires no connectivity at all -- it must be capable of relying only on its own sensors. That is what everyone is pushing towards and that is the situation I am talking about. If we want to limit self-driving cars ONLY to well-defined regions, and require that they ALWAYS be capable of manual operation, then that is a different situations. But, again, that is not where the push is trying to go.

My neighbor is a deputy sheriff and I was speaking to him a couple of years ago about this very thing (but for a different reason). He mentioned that one concern they always have is making a stop in a location where there is neither radio nor cell coverage. Normally they contact the dispatcher as they are making the stop or at least before getting out of the patrol car. One of the reasons is so that the dispatcher is aware that a stop is being made and if they don't check back in within a certain time window they will assume something bad has happened and will send assistance. But if the stop is going to be made in a location that prevents that communication, then the officer either has to see if they can just follow the car to an area that has communications, pull the car over in the blind with no communications, or forego the stop altogether. When I asked how often that happens he said that it happens many times a day across the department and, for him personally, a few times a week. Now, in heavily populated areas such as "northeast US", that may not happen anywhere near as often as near "Larkspur, CO".

Even for dispatched calls, the dispatcher often lacks full information. The person calling in the accident often isn't at the scene (no connectivity there) and had to get to someplace that does -- and in this case it is usually a passerby that also has only limited information about the accident. So all the dispatcher can do is give the officer a general location and the officer has to just drive along the indicated road until they come across the scene -- at which point they may not be able to contact the dispatcher to even inform them that they have arrived on scene. This is one of the reasons that they often dispatch to vehicles to a reported accident, so that one can drive to an area with comms coverage to request the proper response units.

Heck, even with perfect comms confusion happens even in the middle of the city. When my Jeep caught fire I called it in via cell phone and told the 911 operator exactly where I was -- on the flyover ramp from southbound I-25 to eastbound I-225. The police and fire trucks both had trouble finding me (despite a rather good amount of black smoke rising up from my flaming truck) because the dispatcher didn't give them anything approaching decent information. Part of that was my fault because I had overlooked that fact that even though I-225 runs due east-west at that point (and it ends at I-25, so it only goes east from there), it is an odd-numbered interstate which means that, officially, it is considered to go north-south so, technically, I was getting on northbound I-225. The dispatcher told them southbound I-225 so the police and fire were two levels below me hunting around down in the concrete canyon.

 

This is actually an interesting question that I had a long and fascinating discussion with our recent NSA Visiting Professor. I mentioned that it seemed to me that all of these new technologies always bolt on security as an afterthought and never seriously design it in from the beginning. He responded that bolting it on after the fact is the way it should be (or perhaps a more accurate description is that it's the way it has to be). That set me back on my heels (given his background) but over the course of the discussion he pretty much swayed me to his side -- or at least convinced me that he had a very valid view. He pointed out system after system that has attempted to build security into it from the beginning and noted that none of them have ever seen the light of day, including a secure operating system that the NSA has been trying to develop for a couple of decades now. Systems that do see the light of day usually have a relatively basic security foundation and then, as the issues are uncovered, they are addressed through patches and upgrades.

That's sort of a revision of history at NSA. The reason many NSA systems are in the dustbin of history is that they were horrible at the user level, not transparent to the operators and almost impossible to configure unless you had a PhD in computer security. Even old gear that I worked on was at times unsuitable because the security implementation made it almost impossible to use correctly out of a lab.

NESTOR was a classic example:
https://en.wikipedia.org/wiki/NESTOR_(encryption)

 

That's sort of a revision of history at NSA. The reason many NSA systems are in the dustbin of history is that they were horrible at the user level, not transparent to the operators and almost impossible to configure unless you had a PhD in computer security. Even old gear that I worked on was at times unsuitable because the security implementation made it almost impossible to use correctly out of a lab.

NESTOR was a classic example:
https://en.wikipedia.org/wiki/NESTOR_(encryption)

I don't know to what degree it is a revision as much as a different aspect. I can easily see that one of the reasons for the problems you mention is that the focus was so much on security that all else had to give way.

 

I don't know to what degree it is a revision as much as a different aspect. I can easily see that one of the reasons for the problems you mention is that the focus was so much on security that all else had to give way.

The different aspect is they also try to exploit weaknesses in systems as part of their charter and have incentives to not disclose those weaknesses to the public.
There are plenty of secure systems like PGP that NSA tried to neuter when they had an almost total stranglehold on the public use of commercial strong encryption. I'll just say the level of public trust is low.

Today, with social media, the need to crack encryption has lessened.
https://www.theonion.com/cias-facebook-program-dramatically-cut-agencys-costs-1819594988

 

https://www.sun-sentinel.com/local/...0190801-zokyrcqxz5cn7jpjeenaoud6iy-story.html

A South Florida family wouldn’t be grieving the loss of their husband and father if it weren’t for Tesla’s Autopilot feature, according to a lawsuit filed Thursday in Palm Beach County.

Four months ago, 50-year-old Jeremy Banner was killed in west Delray when a tractor-trailer pulled out in front of his bright red Tesla Model 3.

About 10 seconds before the crash, Banner engaged the Autopilot system, investigators found.

Less than eight seconds before the collision, his hands weren’t detected on the steering wheel, which would have prompted warnings from the car’s automated system, the National Transportation Safety Board said in a preliminary report.

 

I'm with @spinnaker . I think these little guys will win the race.

Unlike self driving cars the accidents kind of clean themselves up (or down).
5G will make a lot of difficult communication problems much easier.

 

I'm with @spinnaker . I think these little guys will win the race.

Unlike self driving cars the accidents kind of clean themselves up (or down).
5G will make a lot of difficult communication problems much easier.

You mean we will live in a Jetson's cartoon?

 

Lots of research going on for the self driving car. Some happening right here in Pittsburgh. My co-worker's nephew is the wrench for one of the projects here.

Will we see a real self driving car in our life time? I just don't see it outside of experimentation under controlled conditions. When humans drive it is amazing the number of seemingly simple decisions we make that would be extremely complex for a computer to perform.

Encounters with flagmen or police directing traffic for example. The police might direct you into a lane that might have normally been an on coming lane. Or perhaps a road that had an unscheduled closure. A sign is posted. The human reads the sign and makes a simple adjustment to his route.

I just can't see the computer being able to recognize these cues and then be able to react to them appropriately.

Perhaps it would work in a closed system like a campus but I just don't see the self driving car going out into the real world. At least not for some time to come.

A self driving flying car might actually be easier to implement.

If you knew the truth, you'd be dropping square bricks out your backside.

 

I'm with @spinnaker . I think these little guys will win the race.

Unlike self driving cars the accidents kind of clean themselves up (or down).
5G will make a lot of difficult communication problems much easier.

Bandwidth isn't always the answer. Sometimes wavelength will save your life, or the lives of those you love, when bandwidth will get them killed.

 

Bandwidth isn't always the answer. Sometimes wavelength will save your life, or the lives of those you love, when bandwidth will get them killed.

Not sure what your driving at, but doesn't 5G use multiple wavelengths?

 

It's currently level 3 autonomous driving on one route but pretty cool.
https://gizmodo.com/ups-has-been-delivering-cargo-in-self-driving-trucks-fo-1837272680

 

Johnny Otto Capt.
https://www.newscientist.com/articl...b-the-flight-controls-gets-its-plane-licence/

Unlike a traditional autopilot, the ROBOpilot Unmanned Aircraft Conversion System literally takes the controls, pressing on foot pedals and handling the yoke using robotic arms. It reads the dials and meters with a computer vision system.

Spoiler

 

https://www.thedrive.com/news/30074...s-with-teslas-driverless-smart-summon-feature

 

https://www.nbcnews.com/tech/tech-n...t-recognize-n1079281?cid=referral_taboolafeed

Uber told the NTSB that it "has since modified its programming to include jaywalkers among its recognized objects," but other concerns were also expressed in NTSB's report.

 

https://www.nbcnews.com/tech/tech-n...t-recognize-n1079281?cid=referral_taboolafeed

The family took the money because IMO the fact Elaine Herzberg was jaywalking was the root cause as to why this happened. I can't predict when some random drug addicted homeless person will jaywalk on a dark street so it's not any great revelation that the Uber driving system didn't.

 

https://www.ntsb.gov/news/events/Documents/2020-HWY18FH011-BMG-abstract.pdf

On March 23, 2018, at 9:27 a.m., a 2017 Tesla Model X P100D electric-powered sport utility vehicle (SUV), occupied by a 38-year-old male driver, was traveling south on US Highway 101 (US-101) in Mountain View, Santa Clara County, California. At this location, US-101 has six southbound traffic lanes, including a high-occupancy vehicle (HOV) exit lane to State Route 85 (SR-85) southbound on the far left. As the SUV approached the US-101−SR-85 interchange, it was traveling in the lane second from the left, which was an HOV lane for continued travel on US-101. While approaching a paved gore area dividing the main travel lanes of US-101 from the SR-85 left-exit ramp, the SUV moved to the left and entered the gore. The vehicle continued traveling through the gore and struck a damaged and nonoperational crash attenuator at a speed of about 71 mph. The crash attenuator was positioned at the end of a concrete median barrier. As a result of the collision, the SUV rotated counterclockwise and the front body structure separated from the rear of the vehicle. The Tesla was involved in subsequent collisions with two other vehicles, a 2010 Mazda 3 and a 2017 Audi A4.

 

Jeep caught fire …………. on the flyover ramp from southbound I-25 to eastbound I-225.

So now I know who clogged up the road. lol Sorry for the loss of you Jeep.

Almost the same story but with a tractor fire and a grass hay field. The fire truck drove to the cell phone location but could not jump the irrigation ditch. "We don't follow you driving instruction. We just go the way the computer said." Use Hwy60 and follow the black smoke. (not hwy 287) How hard is that?