Self-Drive Cars Invade Frisco, TX

I assume everyone here but me already knew about this.

I had no idea they were coming to Frisco, and wondered why the news was so hyped over this last week. Apparently, I failed to pay attention to the stories, and assumed this was yet another “test” in some out-of-my-area place. I was wrong.

2 Likes

Wow…I didn’t know. Gotta wonder if other cities are OK with this, as I assume the car service might have destinations outside of Frisco?

lol…gotta love the author’s complete non-understanding of North Texas:

Maybe that’s why Mountain View–based startup Drive.ai isn’t taking its robo-cars to the land of the Grand Canyon, but to the wilds of the Lone Star State.

Frisco, Texas, a city of 164,000 people on the northern edge of the Dallas-Fort Worth metro area.

Makes Frisco sound like Pig’s Knuckles, TX, just this side of Nowhere. In reality, it is so built up and crowded, people that moved there 10 years ago to get away from the crowds are moving even father North/East to get away from the crowds again.

4 Likes

My impression is that it will not, but my understanding is that they have yet to share route information.
By my reckoning, this will primarily serve the Dr. Pepper stadium, Jerry World Jr., Ikea, etc. and the surrounding office parks. But I could be waaay off base…

1 Like

Author need a more interesting way of saying Drive.ai snubbed Arizona in favor of Texas, perhaps. :smiley:

1 Like

You’ll notice that the article said that there would be a human driver behind the wheel.

Announced today, the six-month pilot—which will keep human safety operators behind the wheel, ready to grab control if the car gets confused or misbehaves—

If the past is any guide the “human driver” is to provide a fall guy (or girl) if anything goes wrong. The Amazon car that crashed and killed the driver is an example. The accident was still under investigation and Amazon announced that the driver was at fault because he was supposed to take control and avoid the barrier. So the driver is expected to prevent the self driving car from having an accident. And he failed.

Russell

3 Likes

I am all for driver-less cars (eventually). They need to get some of the legale/isurnce/responsibility stuff inroned out, though.

Best approach I have seen discussed is that used for vaccines and vaccine manufactures: Vaccines save lives. A lot of them (oh oh, did I just start another fight?). But sometimes people die from them, clearly wrongful death. But you can’t hold manufacturers responsible for something that is 1) statistically unavoidable and 2) a huge social benfit. So legislation was implemented that created a tax on vaccines that go into a wrongful death payout fund, and the limited number of deaths caused by this clearly necessary function are paid for out of it, essentially a pooled risk approach.

So, same thing with driver-less cars and deaths/accidents cause by them.

2 Likes

Does that mean there will be less “Hurt in a car accident” billboards or more?

1 Like

“I’m Jim Adler, the Texas Hammer!”

lol…one of the best reason yet to advocate for quick adoption of driver-less :slight_smile:

2 Likes

My personal injury attorney friend is trying to get out of personal injury partly because she is aware that she probably won’t have a job someday due to upcoming technology. Some of her cases are pretty horrible and she is looking forward to driverless cars. Not all personal injury attorneys are terrible people. I expect that the sleazy personal injury attorneys will try to delay this technology as much as possible though.

3 Likes

Wouldn’t people still need lawyers if it does go the way of the payout fund?

There will still be a need for some personal injury attorneys but I would expect the need for personal injury attorneys to be much less than it is now.

1 Like

Pity the safety driver - a job that’s worse than the typical security guard. As a security guard, you might be fighting boredom waiting for something to happen against the otherwise a whole lot of nothing, but at least you have engagement with the process. As the driver of a Tesla with “Auto Pilot” engaged (or worse - safety driver in an autonomous test vehicle), you have to pick out a specific something to intervene in against a backdrop of otherwise highly-similar something that you otherwise don’t engage in.

This is why SAE level 2 autonomy is not advertised as anything close to “self-driving” by all OEMs not named ‘Tesla’ - when those systems fail they tend to fail hard. They’re better suited to transparent intervention such as lane assistance, emergency braking, adaptive cruise control where driver engagement is maintained since the level 2/3 automation has some pretty sharp limits.

Level 3 is a reasonable bar to shoot for in the short term for personal motorcars. The system handles its use cases but also doesn’t demand constant driver attention to a process they’re otherwise not fully engaged in. The grace period to handoff is critical here since the driver will need a second or two to get a feel for the situation before assuming control. GM’s Super Cruise feature is close to this level of automation as is Audi’s Traffic Jam Pilot - both are geo-fenced to specific areas and situations where the system has the data or sufficient confidence in its sensors/processing power to handle the situation.

But level 4/5 is where everyone serious about true autonomous cars is aiming. Level 4 treats driver controls essentially as an override-mode and to handle operation outside of the use cases the system is limited to; the driver can otherwise stop paying attention. With Level 5, driver controls are entirely optional and the vehicle can handle driving itself under any reasonable conditions.

Handy graphic on vehicle autonomy levels

With regards to autonomous cars, I’m sure I’ve said it before but I’ll say it again - look to the aviation model of operation for how they’re apt to be used:

  1. The first iterations will likely be prohibitively expensive for most people to own
  2. With the summon-ride-disembark use model the window to impress others narrows to the point that the value of vehicle itself as a status symbol disappears
  3. (1) & (2) will strongly encourage mobility-as-service usage; thus the first generations will be sold/leased to fleet operators
  4. Initial public distrust of the technology might lead to regulatory regime that further discourages private ownership:
    • Vehicles require individual ceritifcation upon assembly with periodic re-certification
    • Maintenance must be done by certified technicians/mechanics; certain major maintenance likely requiring an inspection
    • Only type-specified components can be used for maintenance
    • Chain-of-custody requirements might be imposed when it comes to access to safety-critical systems
  5. I expect liability under these conditions will be on the fleet operator.
  6. Once autonomous vehicles get cheap enough for individuals to own - if ever - the structure of the economy may have changed to the point that it’s still impractical for most with garages disappearing, huge parking lots directly adjacent to retail and workplaces in-filled, and recovery of so much other space dedicated to an assumption of a 0.91:1 vehicle:adult ratio beginning to be recovered for other purposes.
1 Like

I should have clarified that in my (uneducated) opinion, the game changer for safety won’t actually be driverless cars, it will be vehicle-to-vehicle communication.

1 Like

Reliable, secure V2V comms will surely extend the sphere of awareness for all vehicles. If you know a lane is closed 2 miles up or you can plan around it. If you know that rush hour traffic on the highway has dropped by 20MPH <1/4 mile ahead around a bend you can let off the accelerator.

But it won’t deal with the realities of human drivers that don’t like to follow rules not to their immediate benefit, who drive unsafely, are easily-distracted, and can’t drive consistently. Scores of safety innovations have made cars safer in the event of a collision (seat belts, seat belt tensioners, energy-absorbing panels, airbags, crumple zones, headrests) and made accidents easier to avoid (radial tires, antilock brakes, emergency braking, collision warning), but the human factor remains a stubborn and prominent element in collisions.

Tables 9 ( a ), 9 ( b ), and 9 ( c ) tell the grim tale of the bulk of causation:

  • ~2M driver causes
  • ~44k vehicle deficiencies
  • ~53k roadway/weather causes
1 Like

Hmmmm…Let’s put a tax/fee/surcharge/etc. on “Come sue with us” advertising. Could fund tort reform efforts.

Concerning autonomous cars: If it comms over the ether, it can be jammed. And eventually someone will find a way to hack it. There will be few, if any exceptions.

Some of the comments here remind me of a short conversation with a younger member about autonomous cars. Conversation grinds to a impasse when Walter made the comment “I know how programmers write code. No way would I trust my well being it.”

Now turn off/put down your bleeping mobile gadgetry, quit efffing with your infotainment system and drive your dam car. QED without distractions.

2 Likes

There is an old anime about driverless cars.

1 Like

I wonder what the aftermarket car scene will be like.

A: Nice ride. Does it park itself?

B: Thanks, Yup, it also alerts me when it goes to get itself washed when needed and has a stance feature, as well as an autonomous drift mode.

How many exaflops you got under the hood of that thing?

2 Likes

Heck, if it has I/O it can be hacked. Since the hypothetical 1960s model of “centralized traffic controller” hasn’t materialized in the last ~decade’s worth of autonomous vehicle concepts and V2V comms is expected to be treated as supplementary information it’s not apt to be as potentially devastating as realtime hacking of WiFi-controlled drones.

With all the information out there on how to take command of vehicles via ODB2 dongles, onboard assistance, and even TPMS systems, the dearth of such attacks beyond proof-of-concept suggests that it’s a combination of too difficult and people aren’t quite so evil as to actually do this in significant numbers.

You trust your well-being to programmers literally all the time in modern society, like it or not. Based on the sketchy habits of my fellow drivers and the reasonably-good preliminary results we’re seeing with autonomous vehicle trials, I think I’ll continue to be optimistic that safe and reliable autonomous vehicles will arise within the next ~decade.

Like most heady ideals, reality continues to disappoint. Either this is a result of moral failings, or something more deep-seated in human nature where tasks like driving that can be managed as a background process simply fail to captivate enough humans enough of the time to demand full focus.

I’ve seen the first episode and it came to mind since wrangling runaway autonomous vehicles was the central plot device.

1 Like