Ban it
There is no technology that is idiot proof, especially when driving at 60 MPH with hands off the wheel.
The US National Transportation Safety Board (NTSB) has published a preliminary report into last month's fatal crash involving a 2019 Tesla Model S in Texas. The crash happened at approximately 21:07 local time on 17 April this year. Two men entered the car, one in the driver's seat and the other in the front passenger seat ( …
Fundamental problem with any AI or ML system is that it has no conscience, self awareness or self preservation. It doesn't care or know if it kills itself, its operator or any other person or animal in the process.
A donkey pulling a cart is more intelligent than any of this sh!t which is being wilfully set loose on our roads.
The day that computers gain a conscience we are all doomed, though.
"haven't had an eye test for 30 years"
In the US, an eye test is required. I've have been able to renew by mail the last couple of times, but this time I'm required to present myself for a new photo and eye test. Retesting people is pointless if they have a good driving record. I passed my test many years ago and the laws haven't changed all that much in the mean time. The only new ones they pass now are either redundant or are slap-the-forehead common sense items.
When California had a medical marijuana program, to get an official card, you had to give up your driving privileges. Most people just got a note from their doctor and the filth just overlooked that it wasn't in accordance with law. The point is, since many will miss it, if you have a condition that requires you take medication that could/will affect your ability to drive, you may lose your license. I had a friend in high school that suffered from some seizures and lost his license. He was put on some meds to prevent them, but taking those meds means you lose your license.
Besides all of that, driving while impaired due to anything is against the law. Most of the time people call the law "drunk/drink driving", but the actual terminology encompasses any impairment.
This post has been deleted by its author
Ban what? Old people driving cars?
The situation here is:
* Two people get in a car
* Crash spectacularly less than 200metres down the road
* Tesla claim autopilot wasn't engaged, and even if they're lying, NTSB failed to make an exemplar car engage Autopilot on that stretch of road (unless you're then making the case that Tesla quickly pushed an update disabling AP for that specific stretch to cover their tracks, in which case I have a fine range of tinfoil hats you might enjoy).
Based on the evidence, Occams Razor suggests the most likely sequence of events went something like:
Jimbo gets in the car with Dimbo, says "hey, watch this", thinking to show off the spectacular acceleration common to EVs and Teslas in particular. Enables launch control, hurtles 100m down a 30mph residential road and loses control. Neither bothered to put on their seatbelts because hey, they're only taking a quick turn around the block - how much trouble are they going to get into?
I'm more than happy to see flack thrown at Tesla when it's deserved, but the way this played out on a notionally 30mph road looks far more like someone playing the fool as they depart the house.
Apparently it logs everything to HQ, unless you disable it. And if it's a car Tesla send to the press so they can review it, Tesla make it so the reviewer can't disable the logging.
After an episode of Top Gear showed the car "running out of electricity" and being pushed, Tesla looked at their logs and announced that the car hadn't actually run out of juice, they were just saying that because it made better TV. Musk was apparently shocked that Top Gear would stage something entertaining rather than reporting boring facts - he clearly hadn't watched the show enough.
To try to prevent further "misleading" reviews, Tesla made sure that reviewers know that the car has logging enabled.
After an episode of Top Gear showed the car "running out of electricity" and being pushed, Tesla looked at their logs and announced that the car hadn't actually run out of juice, they were just saying that because it made better TV. Musk was apparently shocked that Top Gear would stage something entertaining rather than reporting boring facts - he clearly hadn't watched the show enough.
To try to prevent further "misleading" reviews, Tesla made sure that reviewers know that the car has logging enabled.
To be fair, Top Gear's response was "we were simulating what would have happened if it ran out of juice". Which was a stonking bit of bullshit. I can't recall the last time they pushed a regular car off the track "to simulate what would happen if you drove it until it ran out of petrol".
They also had a US reviewer driving a Model S round a car park in circles running the battery down so they could claim it had "only just" made it to the end of a road trip (which was something like a 150miles, and the review claimed it badly underperformed it's advertised 250mile range).
Whilst Musk is exceptionally full of himself and his products, the world of auto journalism has lost a lot of credibility over it's knives-out approach to electric vehicles. It was stupidly easy to see where Ford and GM were bunging brown envelopes.
One might give Top Gear a bit of credit as an entertainment show, but five years ago there were lots of "serious" pieces written by previously-reputable journalists which were quite literally lies and misrepresentation. It's not entirely unfair that Tesla started fighting back by exposing those reviewers.
Complaining about Tesla's poor build quality is fair game. Stoking the fire of range anxiety by lying about performance should be career-limiting.
No one else would let it go - if you said the new M3 had terrible acceleration and didn't go very fast but then caveated that you hadn't exceeded 2500rpm or gone above 4th gear then BMW would quite reasonably suggest that it comfortably met the advertised performance if you drove it properly.
If the storage has been destroyed how the fuck did they know that??
The dashboard infotainment storage was destroyed but the black box safety data logger was read by the NTSB.
Ps if your car has an airbag it likely has data logging recording the rate of deceleration before the impact, even if there is no other spyware in the car
The dashboard infotainment storage was destroyed but the black box safety data logger was read by the NTSB.
citation needed
What black box? AFAIK from watching various Tesla teardowns and previous issues, like the excessive logging & NVRAM problems, there's only 1 storage. But that's not to say there couldn't be some form of 'dying gasp' that triggers on say, airbag activation and transmits some data back to the mothership.
"although the restraint control module (which records data relating to the vehicle's speed, acceleration, belt status and airbags) was recovered and is being evaluated, NTSB added."
ie the airbag data logger. Almost all airbag systems contain a data logger to record the accelerometer data that led to the decision to fire the airbag. they were added because manufacturers were worried about "the airbag went off and caused the crash" lawsuits
The dashboard infotainment storage was destroyed but the black box safety data logger was read by the NTSB.
citation needed
Try the NTSB report linked in the article.
Tesla cars have an Event Data Recorder module, which is separate from the infotainment and other systems. Like an aircraft black box, it is designed to survive a crash and records the sort of informations that investigators and lawyers would want to see. You can even buy the analysis kit from Tesla: https://edr.tesla.com
"If the storage has been destroyed how the fuck did they know that??"
1. The car sends telemetry to Tesla via cellular data and the initial impacts would likely have triggered the transmission of event data, before the fire took hold;
2. From the linked NTSB report:
"The fire destroyed the car, including the onboard storage device inside the infotainment console (figure 2). The car’s restraint control module, which can record data associated with vehicle speed, belt status, acceleration, and airbag deployment, was recovered but sustained fire damage."
i.e. the Event Data Recorder module was damaged but not destroyed.
i.e. the Event Data Recorder module was damaged but not destroyed.
So a fun trip down the rabbit hole. In which I discovered it's a 'PSRCM' module. And also discovered it's connected via canbus to a variety of things, including seat weight detectors*. Also found some diagnostic tools that can dump memory, but no free stuff to show a sample decode.
But from the sounds of things, the data may not have been read, ie recovered damaged and taken off for further investigation. Which presumably means there's still a bunch of speculation. Plus or minus correlation with other reports, like the consumer report showing it's possible to have a driverless Tesla. So seems easy to fool the PSRCM by just fastening the seatbelt on an empty seat, but a bit more challenging to fool the weight detector. Unless there's a bug.
Still curious how Tesla would be able to state with confidence what was logged in the module. I guess it's possible to still use a dying gasp to dump from PSRCM to it's main system board and then to the mothership, before the fire destroyed that. Still curious what data it can log. Also discovered there's a bunch of tools that can re-write PSRCM/SRS modules so you don't have to buy a new unit. Or I guess can try to fool insurance investigators, ie alter speed stored from say, 135mph to 35mph, and fudge accelerometer reading to match. Somehow I suspect investigators are wise to that though.
*Apparently so the mass of the driver can be fed into some complex and possibly proprietary algorithm that can adjust the bigness of the bang. Alternatively, with a spot of creativity and an open write channel, be modified to help fight obesity. Check mass, check destination, tell driver to 'Walk you lazy fecker!' and refuse to start.
"but a bit more challenging to fool the weight detector"
Hmmm, how about a sand bag or disc weight from the home gym? Bjorn Nyland used a dumb bell to trick cars he's tested to stay on when they didn't have a "camping" mode by placing it on the driver's seat to simulate somebody sitting in the car. That would let the HVAC keep running while you take a nap in the back.
Any increase in the amount of data collected and sent to others amounts to a worse invasion of privacy. I would argue that it is far less practical to avoid using a car than to avoid using a mobile phone (at least for selected periods of time where we do not want our location to be known), or to avoid sending private data to social media. Right now we can of course avoid using a car that collects such data and sends it to HQ, but it is likely that more and more cars will be doing so until we will soon not have such a choice.
"If the storage has been destroyed how the fuck did they know that??"
Yes, the Tesla's stay linked to HQ all of the time they can. You can opt out but you lose updates, Supercharger access, phone app, etc. They make it very painful to NOT take the constant tracking. Most people are happy as clams to fork over their privacy if they get some nice shiny beads to fondle.
I wish I had a stack of cash to buy some wrecked modern EVs and RvsEng the phone home functions and figure out an off switch. I expect that I could make some coin offering a delete service to high profile people. It's on my list, but I need to pay off the house first and put together a budget to put up a workshop.
For those blaming age - the driver was 59, which I certainly do not consider to be particularly "old" these days. But if you believe that a 59 year old is too old to operate a car safely, then surely they are also incapable of doing many other tasks adequately, and we should set the mandatory retirement (and pension) age accordingly?
If that's the case, at what age would you suggest we revoke a person's driving licence and put them out to pasture?
"at what age would you suggest we revoke a person's driving licence?"
What's age got to do with it? Compulsory competence-based tests every few years, if you fail you get a few weeks to retrain and retake the test, or you lose the affected categories of the licence.
Where's the problem with that?
"Compulsory competence-based tests every few years, if you fail you get a few weeks to retrain and retake the test, or you lose the affected categories of the licence.
Where's the problem with that?"
Cost. There needs to be more data points than just age. Prince Phillip's little crash coupled with his age and physical condition did it for him driving any more. If everybody is required to retake their test, either the government is going to have to foot the bill for loads of unnecessary tests or make people pay private contractors to get tested. Either way it's more tax on driving which is already bad enough. Just add up what it costs all in to own your own car and drive it. The last thing I want is yet another mandatory "fee" I have to pay. It's like the usury fee the post office charges me for a PO box when I've shifted all of my mail/deliveries to the local post office. It costs me money to NOT have them deliver all of it to my house. Not that it would stay delivered where I live.
There is no technology that's idiot proof - that's why thousands of people a year are killed by the actions of the nut behind the wheel.
Autopilot isn't yet marketed as a replacement for paying attention to the road, and neither should it be. But it's use does appear to result in fewer collisions than would otherwise occur.
The sooner we ban people from driving the better.
> There is no technology that is idiot proof, especially when driving at 60 MPH with hands off the wheel
While I'm at least as wary as anyone else when it comes to the current state of self-driving technology, the entire point of this article is that autosteer couldn't be engaged on that particular road, and therefore
The Tesla in question was equipped with Autopilot, which requires both the Traffic Aware Cruise Control and the Autosteer systems to be engaged. The former is a jumped-up cruise control, which deals with acceleration and deceleration while the latter assists with lane keeping. The NTSB showed in tests with an exemplar car that the latter also could not be engaged on that part of the road.
So I'm just saying, it might be worth actually reading the article before grabbing a pitchfork to wave at the nearest sentient toaster.
Beyond that, Ars Technica has a much more detailed article, which actually reveals a more troubling possibility (with the worrying part emphasised).
https://arstechnica.com/cars/2021/05/ntsb-finds-no-reason-to-suspect-autopilot-in-fatal-tesla-crash/
As for why the driver was not found in his seat, one troubling possibility is that the front door was inoperable or obstructed and the driver died while trying to escape from the rear of the Model S. Unlike most cars, Tesla uses IP-based electronic door locks that fail if the car loses power (as it would have in this crash). Although the front door handles will continue to work in an emergency that cuts power to the car, under such conditions the rear doors of a Model S can only be opened using a plastic tab found in the rear footwell.
For me, the fact that in the event of a power-cutting disaster you can only open the rear doors by rummaging around in the footwell to find a release tab hidden under the carpet is an absolute NOPE.
Isn't that plastic tab in the footwell located right on top of the (on fire) battery?
In this case, it may have been a case of 'check under the tree'.
The more indepth report will hopefully explain the sequence of events better, ie if the driver fooled safety features, and which features were active. I'm wondering if it may have been doing the fooling, setting a cruise control speed, then hopping out of the driver's seat to demonstrate it's auto-da-fe features. In which case it wouldn't have needed lane assist, and simply launched itself into oblivion. But that would still leave the question as to why collision avoidance didn't kick in and slow the car.
"But that would still leave the question as to why collision avoidance didn't kick in and slow the car."
The same question could be asked about the car on a California freeway that definitely was on Autopilot slamming into a wrecked semi truck also hitting a person that stopped to help the truck driver. That sort of thing has happened a few times now.
"For me, the fact that in the event of a power-cutting disaster you can only open the rear doors by rummaging around in the footwell to find a release tab hidden under the carpet is an absolute NOPE."
The doors on many cars are an issue. I think the drag savings is a bit dubious and a big trade for being able to yank on a damaged door by the handle that might let somebody in to rescue the occupants. Tesla doors are also known for freezing shut and the opening solenoid punching through. That solenoid has very little leverage so it has to be very powerful (and dangerous). Adding non-breakable glass would be a big safety problem too. It takes time to get out the Jaws of Life and if the battery is on fire, being burned to death in your own car (or suffocated) is a big downer. This is assuming that whoever is trying to help has the tools firefighters have rather than the simple glass breaker that everybody should have within reach inside their car.
I gather there is a driver-present sensor and I gather it can be defeated, if committing suicide and claiming your life insurance (if you see what I mean, and, hypothetically and for instance), is your plan. Didn't some researchers manage it the other day, not these ones I hope... A sack of potatoes in the driving seat would be suspicious... but there are some trick crime fiction stories where the suspicious object is made of ice... how about an ice mannequin?
Tesla cars only check at the point of activation, and deactivate only if the seat belt is removed or door opened, or no torque on the steering wheel after a period of time. So put the seat belt in, sit on the seat and seat belt, enable autopilot, move off seat, start the car going (tap the accelerator). Have some touque on the steering wheel now and then.
This post has been deleted by its author
In a previous case reported here they didn't cooperate with the NHTSA:
https://www.theregister.com/2020/02/26/tesla_apple_death/
"Members of the National Highway Traffic Safety Administration (NHTSA), questioned during the hearing, said Tesla snubbed requests to describe how its cars were designed to operate under specific conditions, and that Elon Musk's engineers did not intend to take any actions regarding NHTSA’s recommendations in its previous safety reports."
Has Tesla's policy changed? Do they always cooperate when the NTSB get involved? Or is it on a case-by-case basis?
Very different agencies. The NHTSA is in charge of highways and safety measures (speed limits, crash test standards etc) and is very slow and bureaucratic and does "dialogues-with-stackholders".
The NTSB is used to beating the heads of $bn defense companies and robber baron railroads. A much tougher bunch
What on earth are you babbling about? The NTSB is notorious for not liking to blame pilots. As with the 737-MAX, where despite the hype about systems and sensors, the fact is that the same issue happened many times, and only twice were the pilots - the particularly poorly paid and inexperienced pilots, in those cases - unable to correct it.
The simple truth is that whilst MCAS wasn't working properly, in every case where the pilots were up to the standard we would expect from airlines, it didn't cause a crash.
I would feel much safer flying if I thought that issue was being addressed as thoroughly as the MCAS issues.
So autosteer needs white lines painted on the road to work then. Might as well have a slot in the road to recharge the car and steer it at the same time. IE a big Scale Scalextric.
BTW what happens if two driverless cars meet each other on a single track road?
So many options:
They both go "safe and stop". refusing to go in any direction.
They both drive to the nearside into the verge/bank/ditch/hedge
One or both goes into a dither and whilst it is thinking about what to do drives head-on into the other.
One or both have figured out where the last passing place is and which is nearest so they can back up to it.
I am sure there are plenty more scenarios, all leading to a conclusion of the journey somehow.....
Whilst the vehicle will have been chopped up by the emergency crews there does appear to be an awful lot of damage. You do occasionally get fires with accidents but given the number of petrol cars on the roads, the propensity for a Tesla crash to end with a fire does appear higher. Are the batteries in the equivalent of a fuel tank that is outside of the cabin? Is it that the heat generated by a battery fire is so great that in sort of firewall or containment is just ineffective?
Electricity is a cruel and unforgiving thing at times.
At the power densities in an EV or Plug-in Hybrid pack, a dead short can vaporize cabling/metal, and that's even before you get into the flammable electrolytes.
Even a normal lead acid battery can do the same, but for a shorter time, and it's why some manufacturers have explosive disconnects on remote mount batteries with long cable runs.
In a crash, all sorts of "unintended packaging configurations" happen, and some result in a compromised pack, and things get hot, quickly.
Sure, a tank of diesel/gasoline has more overall energy density, but it also very rarely comes with it's own means of catching fire spontaneously.
Unless it's in contact with a catalyst, which happens from time to time.
Linseed oil on a rag combusts if I remember right.
Or if you're really feeling frisky, outdoors, in a VERY well ventilated area, preferably on something stone/foil lined, in a pail of sand, saturate a cotton ball with cyanoacrylate glue.
Icon because that's what happens.
In the end, it's a trade-off, and I think that a mild hybrid/KERS setup is better than full electric for most use cases. Diesel hybrid would be better, but good luck getting anyone to bankroll that idea.
Exactly what I was thinking of, but in a small car with a stick shift.
Something along the lines of a Honda IMA pancake motor coupled to a VW TDI engine, with a good 5 or 6 speed transmission.
Do it all in a Miata sized chassis and it would be a hoot to drive.
Valid question!
The IMA (Integrated Motor Assist) architecture uses a single pancake motor/generator combo unit that's sandwiched between the engine and transmission, always connected to the engine crankshaft and input of the transmission.
Mainly kicks in for added low-end torque, light load cruise and regenerative braking.
Without some form of variable gearing, the engine would not be able to stay in the power band.
It would effectively be like a powerglide automatic transmission stuck in one of the two gears, or never touching the shift lever in a manual. Reduced overall maximum speeds would one side effect.
The early Honda Insight could be ordered with a stick shift transmission or an automatic transmission.
The Toyota system uses a motor and a motor/generator with a differential inside the transmission to simulate a CVT, and I think the other players (Ford/GM/Daimler) use a similar two motor layout.
If I remember right, Hyundai uses a single motor built into the transmission, along with clutches or a torque converter.
A full electric motor or true divorced engine, generator, and motor setup is different, since an electric motor provides maximum torque at stall (zero speed), and is controlled by varying torque demand.
In that case, you design gearing so that at desired max road speed will not spin the motor too fast, and design the motor so that it can provide enough torque to move the vehicle and attain maximum RPM.
Assuming an AC brushless motor, you vary the motor output torque based on current/voltage supplied, and switching frequency.
Yes, I know way too much trivia on this.
>Sure, a tank of diesel/gasoline has more overall energy density, but it also very rarely comes with it's own means of catching fire spontaneously.
Except strangely in supercars which seem to combust everytime any redditor with a camera passes them
Most highways already have perfectly good dashed white lines painted on them, separating the lanes. They should just work with autosteer. There's no need for anything "special".
Recharging the car with a slot in the road has been done, but it's complicated and expensive and really not worth the hassle. Just stop for a coffee and a loo break, and let the supercharger recharge your car while you're doing that.
> BTW what happens if two driverless cars meet each other on a single track road?
No such thing. There's driver assistance, but no fully driverless cars yet. When there is, it's unlikely that they'll be driving down single track roads at least to start with.
"
Just stop for a coffee and a loo break, and let the supercharger recharge your car while you're doing that.
"
Even if there was a battery that was capable of being recharged in a few minutes, just do the arithmetic to see how much power each charging point would have to supply while recharging. And multiply by the number of cars you would want to charge at the same time at any recharging site to see what sort of power each site would need to be supplied with. First I'll assume a car consumes 250Wh per mile (real-world values vary between 200Wh and 500Wh per mile). Next, assume the vehicle is recharged every 100 miles on a journey, and we want the recharge to take no longer than 15 minutes. All very conservative estimates (few people stop for 15 minutes every 100 miles on a long journey)
Total charge needed each 100 miles = 25kWh
Therefore power supply needed to charge in 15 minutes = 100kW
Assume each service station can accomodate 10 cars, then power needed to supply each service station will be 1000kW or 1MW, which is equivalent to a 4000 amp 250V domestic supply. And for many hours during a bank holiday weekend, it is likely that almost every motorway service station would have every charging bay used and so actually be taking the full 1MW of power. It would thus entail making a very significant upgrade to the national grid in order to supply every motorway service station with a 1MW supply.
And if you want to do more miles between charges, or charge in less than 15 minutes, the amount of power needed would increase proportionately.
The first and last days of a typical August bank holiday sees 16.5 million cars in the UK making an average 200 mile journey each. Total power used per car will be at least 50kWh, so total power needed in each of those days will be an *additional* 825 GWh. This is almost the same as the present daily Summertime electricity consumption of the UK, and so would need the national grid to be able to deliver twice as much electricity if all cars become electric.
It's why I think the only way that electric cars will really success is if there is a mandated standard battery format and these are quickly exchangeable. Sure, there'll be a few duff batteries occasionally but this way the driver can quickly change the battery pack and the returned (empty) battery back can be recharged and checked almost at leisure.
"It's why I think the only way that electric cars will really success is if there is a mandated standard battery format and these are quickly exchangeable."
Even Tesla tried that. It doesn't work. The biggest impediment fight now is that there isn't the battery manufacturing capacity to build droves of spare packs to have on standby.
I didn't say that it was easy, or thay exchangeable battery packs have anywhere near good enough energy density, but it's about the only way that such vehicles will be realistically feasible.
I'd love electric vehicles to be successful, but right now the technology is barely there (particularly electricy storage) but the power generation and infrastructure just does not exist and that kind of thing takes a very long time to build up.
Wow, make assumptions much?
Most new EVs have over 200 miles of range if they are worth buying. The vast majority of people will do a bulk of their charging at home and a few at work. If you don't have off-street parking, move or don't buy an EV.
The peak charging rate for EVs isn't maintained for a long period of time and is only achievable if the battery's charge state is within a certain bracket the the battery temperature is suitable. Charging stations are set up to limit total draw to what the provisioning is good for. Some new stations are including battery storage that is trickle charged so the charging point can offer charging rates greater than what can be supplied from the grid for a certain period of time. This might become the norm for many stations. It would also be a great use for retired EV battery packs.
Some newer EVs will charge from 20-80% in 20 minutes with a suitable charger (big). I've been tracking my long trips for a couple of years and a basic stop (petrol car) is 20 minutes. A stop with a meal is at least 45 minutes. In an EV with 250 miles of range, I'd likely stop at around the 200 mile mark if I could. That's 3 plus hours of driving and a good time for a break. Stopping every 100 miles would be too soon if I was on a motorway. If it was B roads, it would be long past time to stop.
Most people will be charging on off-peak EV tariffs in the wee hours when demand is low to get the best prices. There may be some issues for a while on bank holidays, but using slower charging might be a good trade off for lower driving costs the other 50 weeks of the year. Taking the train might be a good alternative for a really long trip if you don't plan on stopping on a whim here and there but just want to get from A to B directly.
I was just watching a study done by a YouTuber that marched down the math on if the US switched to EVs overnight. The expected capacity increase to meet that demand would be an additional 30% (averaged). That was less than what's already been put in place to supply all of the air cons in private homes since they became more prevalent. The fact is that the switch won't be overnight, it will take decades and as more people take trips on bank holidays and will be paying to charge their cars, the more business will make that happen. For people to charge at home, it's like adding a blow dryer or a kettle to the houses usage. We don't worry about the grid melting down when we switch on the kettle yet the grid operators can see when all of the kettles going on after a popular TV show ends or at the end of a popular football match. They account for those occurrences in their forecasts so they have supply ready to go. Robert Llewelyn had an interview with the UK grid operator about EVs and they are looking forward to seeing more of them. There is a huge amount of capacity at night that they'd like to see filled up. With all of the wind power, there are times when they have to tell turbines to switch off as there is too much. One goal is to be able to tell EVs that rates are cheap and to start charging instead of switching the turbines off.
"two driverless cars meet each other on a single track road?"
Already tried tested and proven: CSMA/CD.
Cars sense multiple access, collisions dodged.
On detecting another driverless car in the same space, the two cars read each others licence plate/VIN number/etc and the highest one backs off for a pseudo random interval while the other one has another go at getting through.
I'll get my coat, thanks...
"if they are dumber than car drivers I'm in big trouble"
Dumber than human drivers? Yep they certainly are. On the other hand, computers don't drink and they can safely text while driving. So maybe it's not as bad as it seems.
In Trouble? Yeah, most likely. Might be time to start planning your move to a very small island. One with no roads.
Based on the way a number of people drive on the road, I would say when push comes to shove, the self drive vehicles will be safer to be around if you're on a motorbike.
Being that I've been in several motorcycle accidents where I was on the motorbike.
* Hit from behind whilst stationary at traffic lights and ending up sprawled in the middle of the junction (Didn't see you)
* Pulled out on from a side road - wrote off the motorbike (didn't see you), driver reckoned I was doing 70mph in a 30mph where there was a red light 20 metres away just behind the side road he pulled out of
* U-Turn on main road with traffic (didn't see you)
Also, lane changes on motorways with no indication and various other near misses including doing a superman impression at 70mph due to a trail of oil/diesel on a wet motorway.
Love riding motorbikes, don't ride them anymore as the roads are too dangerous now
Seconded. Also a biker, and it's easy to spot when someone on front is messing with the radio/satnav/eating a bowl of cereal, as the car drifts inexorably towards the lane boundary followed by a rapid overcorrection - almost certainly without checking the other mirror first, so any motorcyclist that misses the telltale signs and tries to pass is a gonner.
That said, lane assist seems fairly hit and miss. Definitely underwhelmed with the version in the Aud e-tron, although I test drove a VW ID4 a few weeks back and - despite being the same manufacturer and only a year apart - it seemed to work a lot better. Still wouldn't rely on either in the UK, too many bad or missing lane markers.
I too have given up motorcycling. I kept one bike for tinkering and restoration but it gets ridden rarely.
We have a family Japanese car that does the lane driving thing and the auto cruise control and it works very well. I've done a few long journeys with that helping and I am certain I'm feeling much less tired by the driving.
The lane keeping thing does come up and say its for dual carriageways and motorways only. Though it does work on other roads.
I had one that ran into me from behind (at slow speed) while I was stopped at a light.
They lifted my back end off the ground and my plate got hooked into their bumper so I couldn't ride off.
Then just sat there looking at me through the windscreen while I tried to unhook myself.
I used to have a VW MkII camper - because it couldn't do anything interesting I could drive it all the way from the top of Scotland to Nr London and step out fresh as a daisy. Would have been nice to do 70 up hills and not the crawler lane but because it was so limited you couldn't get stressed about anything.
I seem to remember that there was a study in the UK about the effects of not repainting the lines on roads in cities and large towns. See https://www.theguardian.com/commentisfree/2016/feb/04/removal-road-markings-safer-fewer-accidents-drivers
They came to the conclusion that it increased safety, because all the drivers drove slower and took more care. To me, this is a crazy conclusion, especially when people who visit towns infrequently have to do so.
As I sometimes drive in unfamiliar places, not having hints painted on the road about whether the feft lane is a left-turn only, and losing stop/give way markings on roads, this whole idea fills me with horror.
I've half left the rat-race, and live in a rural town, and when I do go into more heavily populated areas, I'm horrified by the four and sometimes 5 lane roundabouts that appear to be in favor with road designers (examples being the A34/M4 interchange just north of Newbury, the M42/A556/M6(toll), and it seems the M5/A358 in Taunton and I'm sure there are many, many more), where the thought of no road markings fills with dread.
I have no idea how autonomous cars are meant to cope with no road markings.
"Might be time to start planning your move to a very small island. One with no roads."
Sark? It does have roads, sort of, but only tractors use them. Cars are banned. Not sure about motorbikes. Although at only 2.1 Sq. Miles, you can walk everywhere in a short while :-)
Can't talk about everyone else's, but lane-keeping and adaptive cruise control in my car generally behave very nicely towards cyclists or motorcyclists, in that they are considered the same thing as a car, i.e. the bike occupies the lane. By comparison, most human drivers will attempt to overtake a bike within the same lane, or only partially switching lane. I suspect that auto-pilot cars may easily turn out to be safer for bikes than human-driven cars.
Seems like the local investigators might have been a bit quick to come to a conclusion that the driver's seat was unoccupied at the moment of impact.
Could the driver have not gone into the back after the accident because the front doors couldn't be opened for some reason? Damage, proximity to fire, remaining locked?
It's that last option that is most scary!
As someone who was in a car years ago as a passenger, doing 70mph on the motorway, that then hit stationary traffic at full speed (driver had nodded off, I hadn't noticed till it was too late!).
I can attest that I was not only conscious, but fully aware of what had just happened and able to think straight. My only initial issue was ringing ears, I was essentially deaf for a couple of hours or so!
At 70mph, you have my sympathies. On the other hand, hitting a tree at 70mph is basically an immovable object. If you hit another car, it will have moved, absorbing some amount of the impact, plus both colliding objects have energy absorbent crumple zones, something a tree doesn't have. Then again, at 70mph, I wonder how much energy difference there is between hitting a stationary car or a sold, immovable tree.
This post has been deleted by its author
The actual text from the NTSB report was:
"Footage from the owner’s home security camera shows the owner entering the car’s driver’s seat and the passenger entering the front passenger seat. The car leaves and travels about 550 feet before departing the road on a curve, driving over the curb, and hitting a drainage culvert, a raised manhole, and a tree."
"Could the driver have not gone into the back after the accident?"
I don't know about Teslas I've literally never seen one. Most folks around here seem to own either cheap small cars or 4wd vehicles with lots of cargo space. But getting into the back seat from the driver's seat in most modern vehicles requires quite a bit of acrobatics. Not impossible for a sixty year old. I did the reverse (back seat to front) a few times when my front door locks froze. But it wasn't especially easy.
I expect that the official accident reports when we see them in a few years will address the question of how these guys managed to keep the vehicle accelerating if no one was in the driver's seat. Having spent a few Saturday nights in Texas I can imagine a lot of scenarios that might seem improbable to someone who hasn't been there. I wouldn't consider two large rocks -- one on the seat and one on the accelerator -- to be unlikely. But there is surveillance video. Surely that sort of stuff would show up?
I reckon we'll just have to wait for the reports.
Mike Hawthorn, the UK's first F1 world champion, was killed when his car skidded off the A3 Guildford bypass and collided with a tree. No setbelt, his body was in the back seat of the car
http://www.open.ac.uk/Arts/history-from-police-archives/RB1/Pt3/pt3Hawthorn.html
I was rear-ended on the motorway when my vehicle was stationary (a pedestrian ran across the carriageway about 5 vehicles in front and I was the last driver to stop). The driver of a car about 50 metres behind was on the phone and "didn't see that the traffic had stopped". Even though he applied the brakes, he hit my car at about 50mph. I was wearing a seat belt, but had a major concussion caused by my head hitting the back window (my seat collapsed). I'm fairly sure that in a major accident the driver and passenger could finish up anywhere, particularly if they are not wearing a seatbelt.
In your case, the impact caused a sudden and massive acceleration, so you were pushed back. The Tesla hit a tree, a sudden and massive deceleration, so they would go forwards. I can understand that stran ge things can happen in crashes as to where people or objects end up in some odd places, but the angle of the vehicle and direction of travel should allow investigators to back track where the "loose" objects were before the impact. I can' imagine a scenario where a body would end up in the back seat after a forward facing impact. But then I'm not an accident investigator and the last time I studied this sort of physics and Newtons laws of motion was in school about 40 years ago.
"The Tesla hit a tree, a sudden and massive deceleration, so they would go forwards."
Assuming the vector was all in forward direction. A big lurch to the side is more dangerous and that's when people can be forced sideways out of their seat belts. Sometimes that's exacerbated through a big push from the air bags.
"But getting into the back seat from the driver's seat in most modern vehicles requires quite a bit of acrobatics. Not impossible for a sixty year old. I did the reverse (back seat to front) a few times when my front door locks froze. But it wasn't especially easy."
When there are flames licking around the outside of your vehicle, many seemingly impossible feats of agility become surprisingly easy. Opening twisted/deformed doors, not so much.
Many people seem to be assuming the driver was deliberately responsible for the high speed and collision but drive-by-wire throttle control has previous when it comes to causing accidents - I would like to think that, at 59 and 69, the occupants had gotten past the "boy racer" stage and either the autothrottle went wrong or the driver accidentally hit the wrong pedal (preferably the latter!) Other than a suicide pact or murder/suicide plot, there are not that many reasons for people to try to do 60 on a residential road.
"But getting into the back seat from the driver's seat in most modern vehicles requires quite a bit of acrobatics. "
If the car is on fire with the front doors jammed, you will find it a bit less of a problem to get into the back seat from the front while looking for a way out. The Ars Technica article pointed out that the rear doors' manual release is under the carpet (just above the burning battery pack). If you haven't read the owner's manual cover to cover, you may not know that. Without power, the door handles do F-all in the rear. Now, getting back into the front might be impeded (smoke/fire) and trying to find something to break a window might be impossible. I have a seat belt cutter/window breaking tool within reach in my car. I worked for a while as an EMT (thankless low paying job) and that made me get a good one. Very few people have them. I have to wonder if Elon is going to get away with non-breakable glass in his toy truck.
"trying to find something to break a window might be impossible. "
In the (former) UK, Aldi or Lidl frequently have a basic car window breaker/seat belt cutter at very reasonable prices, can't comment on quality but it's probably decent. I bought 2 not long ago, I wanted one for home too (to break windows for emergency exit in emergency). Probably available elsewhere too, looks like a tenner at Amazon gets you two at the moment.
"Seems like the local investigators might have been a bit quick to come to a conclusion that the driver's seat was unoccupied at the moment of impact."
As I pointed out at the time, they weren't investigators, they were small-town cops/first-responders. They came up with their usual sensational bollocks. People have been wrongly jailed for years over other cases of the same thing, where occupants of a crashed vehicle were thrown from their seats.
A number of fatal crashes have been attributed to over-reliance on vehicle automation by drivers.
I will not have a car with automation installed. Makes one lazy and tend to enforce the habit of letting your attention wander.
The driver need to be aware at all times of other cars and objects around his/her car.
If you want true automation, get a donkey/horse carriage.
And it's emissions are green
Well, yes and no. One or two donkeys are ok. But when everyone gets them...
And, of course, these days methane emissions are of a concern too.
This post has been deleted by its author
An old quote from Steve Grand: "[W]e used to have intelligent cars; they were called horses. And they used to know stuff that our cars don't know. They used to know where they lived and how to get home and how not to knock people over. Even how to refuel themselves. The amazing thing was that they could even make new copies of themselves. The intelligent car would be like a horse. Something that really enjoyed a good drive and prided itself on not knocking people down."
Automation can be used safely. It should be designed to reduce the workload and allow the driver to spend more time on "important stuff" (like watching what's going on). However, it interacts differently with different people and the way they use the system.
For example, a driver can set (conventional) cruise control to 45 mph and basically drive along a motorway without really caring about what else is going on - they are unlikely to come across anything moving slower or have to do much until they decide to leave the road. Their workload will be low, but it is likely that their concentration will be low as well.
Things may be completely different for another driver using the same system. They may set the cruise control to 70 mph and use the reduction in workload to allow them to scan the traffic so they can spot gaps that can be used to (safely) change lanes to maintain that speed without having to cancel the system. Workload is reduced, but concentration, spacial and situational awareness are improved.
A fully autonomous system needs to be able to cope with an inattentive driver, and that is part of the problem with the various "levels" that are currently recognised within the automotive sector - one of these basically says the car is in full control until it gives the driver the 100mS notice that it won't be much longer...
* Based on some research I was involved with in the early days of autonomous cruise control development.
"A fully autonomous system needs to be able to cope with" a dickhead, or multiples thereof, human beings are inventive and there are no limits to the levels of stupidity some of them are capable of.
If allowing for those levels of Darwin Award winning behaviour is not planned for the auto pilot is not fit for purpose.
Another factor that needs to be considered is that there is a whole class of accidents called "Looked But Failed to See (LBFS)". It's not well understood despite having been the subject of considerable research. LBFS seems typically to happen with experienced drivers on roads that are familiar. The driver, who may well have a spotless safety record -- seems to be normally attentive. But apparently simply fails to process some clearly visible object. Often the object is a motorcycle ... but it can be pretty much anything ... up and including emergency vehicles covered with flashing lights.
Automated hazard detection and notification might well reduce LBFS accidents.
I'm not so sure about the wisdom of letting the vehicle make decisions about the velocity and direction of the vehicle except in carefully controlled situations. It's one thing for the sales folks to run a demonstration of their technical wonder negotiating a well-vetted course of hazards. You can demo pretty much anything if you put enough effort into the demo. It's quite another for something to work in arbitrary situations.
Tell you what. I'll lay out a route on surface streets from the Boston suburbs through downtown Boston to Logan Airport, and we'll have a pool on how far your autonomous vehicle gets before its first accident.
New owner picks up "keys", gets in car and instructs it to take them home.
Car drives across the parking lot and proceeds to road.
It gets no further because it can't work out the "merge algorithm"* for the Bangalore traffic ;-)
* The "rule" seems to be that if there are "n" streams of traffic, where "n" != number of lanes, you push the nose out until there is only enough road left for n-1 streams. At that point you move out, restoring the number of streams to "n".
* The "rule" seems to be that if there are "n" streams of traffic, where "n" != number of lanes
Or
If there are three lanes on the road there will be at least 5 lanes of cars trying to go straight on (or turn left or right, you won't know, they might not have decided yet), then a bus will pull up on the outside of all the lanes and want to turn left across you. The bus is bigger than you and already has more dents. Which probably explains why my Hertz rental car came with a Hertz rental driver.
Perhaps chauffeurs are the real answer to self driving cars.
@Mishak
"....you push the nose out until there is only enough road left for n-1 streams."
Exactly how I learned to negotiate the roundabout at Hyde Park Corner in London in the early '70s when I first got to London.
I was dutifully waiting, giving way to traffic already on the roundabout, as proscribed, then realised after a while that if I continued to do that, I would never get out, because the traffic was unending.
I also adopted the 'black cab' method as described above, which is just slowly edge out until some kind soul let's one in.
My first lesson in driving in London's heavy traffic.
Further to the above post, when I briefly worked for 6 months in Mexico City, driving there was a whole new ball game.
The city is made up (mostly) of alternative streets operating in a one system in opposite directions. These roads seemed to be up to about 6 lanes wide which posed rather a problem if you were in one of the centre lanes and realised you had to make a turn.
If the cross street was going right to left and you wished to turn left from somewhere in the right hand lanes, one attempted to get to the right and positioned one's car in front of the others already waiting at the lights, thus being first away when the lights changed in your favour.
It was even more exciting when trying to turn right and didn't happen to be in the right most lane, in which case, again, one followed local custom, and simply signalled to the right, and started to make your turn from whichever lane you happened to be in. Fortunately, the locals seemed to realise that it was going to happen from time to time, and kindly let one turn without much ado.
I once had the situation, where 3 cars, all in adjacent lanes were turning into the side street all at the same time. Me? I was the sucker out in the 3rd lane.
All 3 cars successfully made the turn.
P.S. that was in 1992, so my memory may be a bit astray about the number of lanes on those roads, but it was lots. It may of course, have changed completely by now, but I fear it is probably worse these days.
>"Looked But Failed to See (LBFS)" ... up and including emergency vehicles covered with flashing lights.
We've had people that have steered in front of a 300ton bright yellow haul truck
It's deceptive how fast they are moving and how close they are sometimes
It often doesn't end well. We put high vis masts on the pickup trucks, but often that just means you know which wheel to dig them out from under.
>"Looked But Failed to See (LBFS)" ... up and including emergency vehicles covered with flashing lights.
There have been a few cases of Teslas that seem to have targeted emergency vehicles covered in flashing lights. A couple of HGV's perpendicular to the Tesla's path and a few that went headlong at speed straight into a previously crashed vehicle.
LBFS seems typically to happen with experienced drivers on roads that are familiar. The driver, who may well have a spotless safety record -- seems to be normally attentive. But apparently simply fails to process some clearly visible object.
I've done this. I once drove through an intersection near home only to see an oncoming vehicle suddenly turn left in front of me. It was a damn close thing. My passenger said I went through a red light. I would have sworn in court that I checked it was green. I try very hard to drive safely, and this was a valuable reminder of my cognitive flaws.
through downtown Boston... Say no more! Boston drivers work hard to earn their homicidal reputation.
"one of these basically says the car is in full control until it gives the driver the 100mS notice that it won't be much longer..."
That's the key problem. The SAE autonomous driving levels system can imply that cars get safer (or at least "better") as the levels go up. The problem is, we're in a deadly valley right now. If the autonomy let's you zone out, you're not capable of racing properly when the car needs you to take over. At lower levels of autonomy, you're still engaged with driving. At higher levels of autonomous driving, the car can handle everything until it gets to the point where it tells the driver "hey, in 1 mile I'm taking exit 98, you'll need to take over from there". If the driver feel asleep in that scenario, the car can pull over and stop safely.
We're clearly not there yet.
"hey, in 1 mile I'm taking exit 98, you'll need to take over from there".
That would work, the real danger with the current systems is that they are totally automatic until they fail and then hand over to you a split second before disaster.
eg car drives into blind bend, there is a truck on the wrong side of the road 10m ahead and it says "abnormal event - reverting to manual , good luck"
IIRC there was a similar problem with autopilot on some turboprop that caused a crash. Wings iced up and autopilot silently compensated for it, right upto the point where the plane could no longer fly and the autopilot switches off and says - "so long, been nice knowing you"
All lane running is fine, as long as you have the auto-detection equipment to instantly close the left lane when someone stops. I was genuinely shocked to find out that our so-called smart motorways don't necessarily have it installed. Without it, it's insanity.
With the technology sorted (and installed), you're not at significantly greater risk than with a hard shoulder. If that seems wrong to you, you're probably one of the vast majority who don't realise how unbelievably dangerous stopping on the hard shoulder is.
Braking radar is a must if I ever buy a new car.
Was on a motorway and realised too late that the 3 lanes of traffic ahead were all stopped and had been stopped for long enough that nobody had brake lights on.
Nice thing about everybody over here driving automatics is that they all have to keep their foot on the brake light pedal while stopped.
"Was on a motorway and realised too late that the 3 lanes of traffic ahead were all stopped and had been stopped for long enough that nobody had brake lights on."
Most times I've unluckily come across a serious slowdown or stop on a motorway, the person "last" in the queue pretty much always has their hazard lights on. And like me, once you confirm the vehicle behind has also noticed and is slowing/stopping and put their hazard lights, the one in from turns theirs off. Anyone at the back, even if no visible vehicle approaching, will almost always keeps their on until they are no longer "tail end Charlie". The fear is that some driver not paying attention is barrelling down at 70MPH+
I have some sympathy with this point of view. But having a car that does have quite a few assistance features, I think there's swings and roundabouts. Initially I was sceptical of even the adaptive cruise control, let alone the lane assist, but I have found in some ways these actually heighten your awareness. Not only are you less fatigued by driving, which is surely a safety enhancement in itself, they are much more alert to your errors than most passengers: the ACC will ensure you never creep too close to the vehicle in front (or fail to fall back enough when someone pulls in in front) and the LA stops you from lane drifts and alerts you when you wobble.
My car will also object loudly (and then brake) if there is no steering input and can somehow tell if you are looking out of the side window or at the radio. There must be some analysis that tells it when I need a coffee and it tells me - it's not a simple timer. And as I have already mentioned in another comment, the front collision avoidance system has already stopped some other motorist injuring me.
So I think the key thing is the differentiation between "assistance" and "automation" - the problem with Tesla's autopilot is the name: only pilots and other enthusiasts know how un-auto an autopilot really is.
I was worried about this when I got a car with adaptive cruise control and lane-keeping assist. It's not enough automation to allow you to just do something else, but it's enough to give your brain significantly less to do. I was concerned that driving would fall under the threshold of too boring to pay attention to.
In practice, it turns out that I use the spare bandwidth to pay a bit more attention to other objects besides the car directly in front and the lanes, e.g. cars on other lanes that look like they may switch without signaling, cars further ahead whose behavior may suggest problems ten seconds from now, things like that.
It's when you get enough automation that you start feeling like you could check your emails - that's where the real danger begins.
"I will not have a car with automation installed."
I use my cruise control all of the time. I wish it was adapted. I find it keeps my speed from creeping up and I also get better mileage. On long trips, it keeps my leg from cramping up.
I don't want anything messing with the steering. The other day I saw a piece of truck tire tread in lanes and eased over in my lane to miss it. The last thing I'd want is the car tugging me back into the middle of the lane. I'll also squeeze over if I can't change lanes to give somebody on the side of the road more room.
> what happens when a road has edge lines at first but then stops having them?
Then it requires the driver to take over.
The system doesn't NEED white lines. We did lane following systems in the 90s just using the correlation of road shape, tire tracks, kerbs, lane edge markers etc. For well built roads with sensible curves it works well.
The Tesla system is deliberately only allowing use on roads with white lines to stop people using it in "unusual" conditions. Driving off-road, into driveways, across fields etc.
>What if the camera gets old and the edge lines are "burned" onto the camera,
This only happens if you hit 88mph, the car time travels back to 1986 and replaces all the CMOS cameras with Videcon tubes
This post has been deleted by its author
Somebody already mentioned it. People used to have this level of “fully automated” back in the days. Horses and donkeys doing the same trips repeatedly knew the drill. Often enough, people dozed off in the wagon only to be awakened at the arrival. Furthermore, those creatures knew when something was wrong and were able to alert “the driver” properly.
When comparing this level of intelligence to this “coke-head invention”, I have to laugh. It's like comparing birds to worms.
On self-driving horses -
In a previous life I got chatting - in a Pub of course - to a retired London drayman. I can't remember if he worked at the Whitbread Brewery in Chiswell Street (Old Joke Alert - "Give me a pint of beer", "Whitbread?", "Yeah, 2 slices") or if it was the Youngs Brewery in Wandsworth, I'd done jobs at both places.
He told me that at each pub he delivered to he'd be given a pint of beer. By the time of his last delivery he'd be 9 sheets to the wind, but could rely on his horses to get him back to the brewery safely.
The past truely is a foreign country.
My grandfather was a roundsman for Lyons' Tea in north east London. His horse, Pimple, knew the round, and Grandpa only had to say "Giddap" as he walked out of the shop he'd just delivered to, and Pimple would make his way unerringly to the next stop. Grandpa never even had to get into the van, but walked alongside it. But, and it was a pretty big "But", at the last stop, Pimple knew it was soon nosebag time, so Grandpa used to have to tie the reins to a lamppost or similar piece of street furniture, otherwise Pimple was off home before Grandpa had returned from the shop. Many was the time Pimple arrived back at the yard some time before Grandpa, who had had to walk several miles to catch up. Then he was issued with a motor van, but that was considerably less reliable than Pimple, and kept breaking down far from base.
Not relevant to the two Texans purging themselves from the gene pool, but I noticed something about me and my car's level 2 automation that's all too familiar in an IT context-
Toyota Safety Sense 2.5 does a very good job helping with cruise control and lane keeping on interstate highways.
On secondary streets where drivers nose out into blind intersections whilst updating their fucking Facebook status, I very deliberately swerve toward the other side of the road, and the safety system provides firm (yet easily overridden) corrective force on the wheel plus some seriously loud beeping.
Constant beeping leads to alert fatigue, which leads to disabling the system. Not sure how many drivers will dig through the menu tree to disable the beep versus just disabling the entire system with a single button press and lose some of the benefits of the automation.