Four hundred miles with Tesla’s autopilot forced me to trust the machine
arstechnica.com
arstechnica.com
"Lane keeping was almost excellent. The car kept itself squarely planted in the middle of the lane without ping-ponging between the edges like some earlier systems did. The car would still sometimes "lunge" at exit ramps when lane markings would fall away to the right or left, though the behavior was never enough to make me grab the wheel and disengage."
So Tesla just has lane keeping and smart cruise control, like everybody else. NHTSA Level 2. But they're way ahead in PR.
Google can handle construction, changed lanes, and offramps, city streets, and follows navigation directions. They're NHTSA Level 3.
Handoff from automation to human in emergency situations is still a big problem. Here's such a case. [1] Car with autopilot rear-ended stopped car on freeway. Why? The driver tapped the brake while approaching the stopped car. This disengaged the autopilot, including the emergency braking function, under the assumption the driver was now in charge. Tesla claims the driver should have understood that, as of firmware version 6.2, the auto-braking function was disabled by any manual braking.
[1] http://arstechnica.com/cars/2016/05/another-driver-says-tesl...
I'm trying to think of a scenario outside of a mechanic's garage that emergency braking would ever be disabled. This should really be mandatory on all cars.
In the lab, opaque textured objects in sufficient lighting are pretty easy to localize by a variety of methods. Reflective objects are quite a lot harder, but there are methods that work. Reflective objects through distortion like rain/fog... harder, but often doable. In low light? Depends.
The weird thing with environment modelling in computer vision is that there are a lot of overlapping methods, but sensor fusion of these methods is rarely attempted. Most robots end up with a pair of laser rangefinders or a pair of cameras or something equally simple to program. Creating a high-reliability automated vehicle that can bypass all the edge cases where these things fail is another thing entirely. You want overlapping, inter-reinforcing models of the world, with algorithms that not only identify items in a well-calibrated system, but calibrate the system in real time based on items identified.
My lab was really excited about replacing stereoscopic computer vision with monocular computer vision for 3D modelling, but the opposite direction is where you want to go if you care about as many 9's of reliability as possible and hardware is cheap. You want a few dozen cameras with fixed orientation, you want parallax laser rangefinders in a variety of bands, you want time of travel LIDAR units, you want a {GPS + 3DoF gyroscope + 3DoF magnetometer + 3 DoF accelerometer + barometer + thermometer}†, you want ultrasonic and radar rangefinders; You want everything you can get. You want enough lenses that if one gets obscured by water, hey, there are others that will pick up the slack, and do it without bothering the driver.
You want to shrink the 'no coverage of object in environment' surface as small as possible.
Nobody seems to respect this on the academic software side. Novelty, sure, but people don't seem to earn grants publishing extra 9's using a mashup of a bunch of pricy sensors. 'Less is more' is not how you get the data needed to make things reliable, it's how you get best-case detection rather than worst-case awareness.
†This combination is now found in quadrotor IMUs. Each individual sensor is inaccurate enough that it would be useless for automation, but combine them to check each other's deficits and you get a remarkably elegant system.
I'm pretty leery of these systems at the moment due to people using them like this. That's just not what they are ready for at the moment and people's unrealistic expectations are leading them down dangerous paths.
But otherwise, even when in manual driving mode, the emergency brake feature should be active and watching for impending collisions.
Also, it was a UI failure not notifying the driver clearly that auto-pilot has disengaged. Even the primitive cruise control in my car has a big green light, that turns off when you tap the break. The more autonomous a car is, the more important it is to clearly notify the driver of control handoff.
BMW's anti-collision system is on even during manual driving.[1] BMW will brake even if the driver is pressing the accelerator. So will Volvo. Top Gear compares the various systems as of 2015.[2] Volvo did best; Mercedes didn't stop soon enough and hit the dummy car. VW's city car only does this up to 18MPH.
Maybe Tesla has a false-alarm problem that BMW and Volvo have overcome.
More to the point, this behavior needs to be standardized, or people shifting from one car to another will do the wrong thing. The National Transportation Safety Board is working on this.[3] In a few years, anti-collision may be as standard as anti-skid is now.
[1] http://www.theverge.com/2015/1/5/7494989/bmw-i3-self-parking... [2] https://www.youtube.com/watch?v=PzHM6PVTjXo [3] http://www.ntsb.gov/safety/safety-recs/recletters/H-15-008-0...
Even though my car is a 6sp manual, the automatic distance keeping is making the commute much more relaxing. I can't recommend it enough.
Funnily enough the radar system is separate to the AEB, which uses cameras. AEB stands for automatic / autonomous emergency braking.
As I'm someone who likes to know how stuff works, especially in cars for familiarity should a real event occur, I did set out to experience how AEB works. I rigged some large but light cardboard boxes, reversed down the road a bit, accelerated quickly to around 40 km/h, kept the throttle pedal steady as though if to drive through them.
All times, although it felt like surely it wouldn't, the car intervened at the very last second. The first time it drove through them by a metre. The next two times the car stopped before impact. System uses 100% brake force engaging the ABS and also putting on the hazard lights for good measure.
I came out highly impressed. It ought to pick up cars better, but I'm not about to actively try that.
An appropriate way to signal it would be distinct from any emergency controls and also show a highly visible signal that you've turned the system off and are in a different "mode"; even a simple separate dashboard switch with a light would be better for it than the brake pedal.
So there's an accident, and I swerve to avoid. I think I'm going to make it but the car isn't certain so it brakes, and it brakes really fucking hard. At this point who is responsible for what happens? Breaking like that fundamentally changes the behaviour of the car. If I had just braked, maybe I stop in time maybe I don't - I chose another option and it is completely valid.
The very idea of that kind of behaviour terrifies me. I'll take a 100% automated car with pleasure, but I have no interest at all in this hybrid control concept. I would turn that shit off on delivery.
I'm having problems understanding your scenario. How exactly does the car breaking prevent you from swerving? The brakes should only stop your car mid-swerve, and you wouldn't collide. The only scenario when speed is an advantage when swerving is when you're trying to needle into an opening in the next lane and you need to match the speed of a 3rd car
It all depends so much on various variables so I find it hard to categorically say "it should always be on."
A deer crossing the road in icy conditions. Its better to hit the deer than lose control.
a large tumble weed blowing across the road with someone tail gating you.
Yes, but the failure mode is not so easy to see prior to use. It's due to how humans react to the new feature itself, not because it's a bad feature, or a misunderstanding about how users 'normally' react. User's may incorrectly anticipate a behavior (such as in this case), or they may expect the feature to fail so unnecessarily take control (cause of some airline disasters).
This is a general problem of computer human interaction. I'd say it falls under the discipline of UX design. Knowing that breaking is an automatic reaction should be a clue that it will make it a poor mechanism to understand user intent. You then know that you should be looking at other mechanisms to more accurately capture user intent.
- Perhaps the auto-breaking feature should have better defaults such err towards enabling rather than err towards disabling.
- Perhaps the control for toggling auto-breaking should not also be the breaking mechanism itself.
- Perhaps context is now relevant, and user input should be interpreted differently when auto-breaking is on and there is an immanent collision.
I think it's important to pay special attention when making any interface change to a well established user behavior, and acknowledge that new behavior leads to unstable and unreliable input. In these cases put extra effort into finding the path of least cognitive burden when the accuracy of user input is the most critical (i.e. immanent collision) even at the expense of adding cognitive burden elsewhere.
Cognitive burden is a critical consideration in all system design from airplanes to cell phones, but tragically is often overlooked.
It is not easy but it should be done. Keeping a car steadily between lanes is not easy either, or take self driving cars that is hard. Besides, research into some of this has already been done with aircraft and pilots.
> It's due to how humans react to the new feature itself, not because it's a bad feature, or a misunderstanding about how users 'normally' react.
It is a bad feature. Feature doesn't exist in a void, it exists only because humans are using it. "Human presses the brakes, what to do next?" is a feature with a human in the loop.
> Cognitive burden is a critical consideration in all system design from airplanes to cell phones, but tragically is often overlooked.
Exactly. Moreover "mode confusion" (human not understanding what mode the aircraft autopilots or other systems are in, and then reacting assuming a different mode) is a common cause of crashes.
There were tools and studies specifically geared towards mode confusion:
http://shemesh.larc.nasa.gov/fm/fm-now-mode-confusion.html
Asiana 214 is a recent example of pilot assuming the autothrottle was on, but it wasn't, so plane crashed. It is a bit like here, the user assumed Tesla would slow down some but then stay in the automatic mode.
I think Mercedes has a pretty comparable system as well, but can't speak from my own driving experience.
http://gfx.dlastudenta.pl/photo_new/d90/b27/097/537/102764
Or like this:
http://img.sadistic.pl/pics/a87df587b218.jpg
Not everyone lives in countries with amazing road infrastructure, the pictures above were taken in a country that is a member of the EU,not some far Russian province.
https://www.google.com/maps/place/US-101,+Mountain+View,+CA+...
Oh that's great news! Now where can I buy a Google car again?
There's a big difference between a general production-ready car and what's basically an internal-use-only car.
It's a great feature, but it's a great feature for boring driving.
I'd much rather have a "boring" driver on the road than deal with aggressive drivers.
I would be so much more likely to make frequent cross country trips to visit family and so much less likely to ever book a flight unless I'm needing to cross water. I wonder how this would hit the airlines in the US.
People fly to save time, and if youre going to a hub city (I.e. Chicago, Los Angeles, New York, Atlanta, etc.) It will likely be cheaper to fly as well, I just booked a flight from San Francisco to Chicago for $148 (driving would easily be $250).
(Which also makes me wonder how you can do that trip in 20 hours. Speeding indeed :)
Additionally, on the intrastate highways, the speed limits are set by local districts so local districts who like making money will often change their speed limit for any highways that go near by them so they can speed trap people.
Last time I drove it took me four days, but I only drove 8 hours a day and took a long way around.
Would be interesting though - I can imagine a Mad Max-like future where roving bandits set traps for autonomous cars being sent on errands by the wealthy elite.
I am not a Tesla owner, so please correct me if I got the cost/mileage of the tires wrong.
Are you sure about this? Tires usually last 30-60k miles.
The bus service from Houston to Miami is a little less okay, but it's because it's 1900 kM and won't have a lot of users (Google suggests public transport would take 34 hours, a train and a couple of buses).
Of course you have to pickup and dropoff at their location, and you don't have a car to use at your destination.
Maybe by itself, but when you factor in how awful air travel is getting, it might have a bigger impact than it otherwise would. I'm planning a flight from Austin to Chicago later this summer and I'm seriously considering driving because my last two experiences flying have been so awful thanks to the TSA.
It's going to be a long time before a Google car can just drive anywhere you tell it to.
Also, errors in street-view are minor annoyance; for a self-driving car, an error in its internal map could be deadly.
And because such a thing would be a huge win for trucking companies I fully expect the US will eventually instrument the entire interstate system to make this happen. (i.e. there is enough money in it to make it happen.)
Fully autonomous local driving is not likely to happen till we have real AI that can understand things like: the guy down the street is unloading a construction vehicle, so make a U-turn and go down a different street.
I believe they say in there that they got highway driving down really early on. They have been focusing on urban driving for the last few years.
https://www.ted.com/talks/chris_urmson_how_a_driverless_car_...
For travel within cities, autonomous electric vehicles will ultimately surpass rail travel. That will happen within several decades.
The cost of car use and ownership will plunge. Software and constantly improving batteries will eat the expense out of the heart of the machines.
What's the cost and usefulness to run high speed rail between two mid-sized Romanian cities that are 60km or 100km apart? Right, it's wildly impractical. Inexpensive, autonomous, electric vehicles however are the future and will be extremely practical for all purposes except very long distance travel between two major cities. No other scenario will justify spending the vast sums of money required to install and upkeep high speed railway - and that cost will only continue to climb with wages and security threats in Europe; meanwhile electric, autonomous vehicles will bring the cost of vehicles down dramatically over time.
If anyone thinks that there is the road capacity for a million self-driving cars to drop passengers during every morning rush hour in Mid/Downtown Manhattan or the City of London, they are deluded.
What's the cost of building from scratch and maintaining year-on-year a road along the same line, including vehicle and operating costs, before subsidies? That is - what would it actually cost you to drive a car if your taxes weren't paying for it?
Hint - it differs for nearly every situation, and even the idea of high speed rail - 125mph or higher - is not necessarily useful in all situations. A decent commuter line with good bus links to every station is potentially more useful than a high-speed line which zips past dozens of towns.
Instead of now where you force yourself to be fully awake, instead you will allow yourself to semi-snooze, until you hit a situation the computer can't handle - and then you are in real trouble.
And a complete stop on the freeway is not exactly safe.
Most who claim they would remain attentive are not aware that they are making false claims, and this likely includes a lot of HN visitors. The better the cars get at remaining on autopilot, the more complacent drivers will become. A really good autopilot is training its users into trusting it to handle everything. If your average commute results in autopilot practically never kicking in, your brain will naturally - regardless of your intent - find something else to occupy your attention. Reading a book in your lap, checking the weather, grooming yourself in the mirror, or even just spacing out and daydreaming.
Frankly I'm of the opinion that self-driving cars are something that will happen in 50+ years. When cities are built with "roads" that are actually tracks; tracks that are designed to handle only these cars and nothing else. I imagine we will not own such cars; they'll be more like subway/metro public transit systems, but with individual pods that hold a small number of people and with hundreds/thousands of small stations to embark/disembark from.
These vehicles should not be allowed until there is no manual takeover. If the vehicle cannot operate at all times on autopilot, it should not be on the road.
This is the only reference I've seen to that http://electrek.co/2016/04/24/tesla-autopilot-probability-ac....
I care a lot about what happens when autopilot disengages.
I truly and honestly believe it is a biological, psychological fact, that autopilot will result in less attentive occupants in the vehicle. We're going to blame "drivers" - who are really now just passengers, expected to sometimes take the wheel - for accidents caused by autopilots disengaging, because the driver "should have been in control". Except that most people are going to be mentally unable to maintain the kind of focus needed to be able to take over a failed autopilot.
Let me put it this way: I believe that today, if my mother were to cause an accident, she would have done something tangible to be deserving of being "at fault". In a future where my mother can be considered at fault for killing someone because she failed to take over an autopilot in under a second... that is not a future I want to experience.
I'm afraid of the moment we begin blaming people for loss of life caused by technology expecting us to interfere when it is innately difficult to do so.
The driver was texting while driving.
I've seen more instances since then.
If you ask me, people can't be trusted to be safe anyway. I'd rather have imperfect autonomous cars than people driving.
No one is advocating actually sleeping at the wheel, I thought my OP was clear about that.
[1] http://www.nytimes.com/2015/09/02/technology/personaltech/go...
And given how drivers have gotten WORSE since smart phones hit the scene, I'm ready for these robots right now. I don't feel safe on the roads at all anymore.
It does, however, automatically adjust your speed to match the lane you're changing to.
At the end of the day, it's all about sensors and software. I'm sure Tesla has better engineers than Fiat Chrysler but I'm also pretty sure it's off the shelf for them and they just bolt it into the car at this point.
Good stuff, though. Can't wait to see full point A to point B navigation.
So...what about those signs where kids spray paint 35 to be 85. Will it zoom to 85?
I've driven roads where the GPS swore the speed limit was 20 clicks slower than the signed speed limit. It is not necessarily safer to drive significantly slower than the posted speed limit, since most traffic will be doing at least that and relative speed differences are a contributor to surprises/crashes/injuries on the road.
The reason for using a density range, is that if the cars are too dense, there is likely traffic, and if they're not dense enough, you'll either have outliers skewing the data, or people driving faster than normal, due to the lack of other cars.
[1] When you're using Google Maps, google crowdsources data on how fast traffic is moving, based on all devices using Maps in that viscinity. You can infer traffic density based on sampling the number of people in a given area providing data. I assume that this data would only get more granular as time goes on, as we add crowdsourced data from autonomous vehicles as well. https://googleblog.blogspot.com/2009/08/bright-side-of-sitti...
Sometimes the signs say 00 in no traffic lol (I think during testing..?)
I can imagine a "smart highway" where the signs broadcast the current speed limit...and hopefully broadcast the current speed.
I agree that at some point cars could learn the rules as well.
If the unsafe condition of the vehicle resulted from a manufacturer's defect, this, again IIRC, generally allows those who experience damages to seek compensation from the manufacturer, but doesn't necessarily relieve the driver of liability (legal liability for any particular harm isn't necessarily exclusive in one party.)
It's nice that you can check your mail or read a news paper in autopilot mode as advertised but doesn't this create the danger of touching the steering wheel, turning autopilot of, while still holding a newspaper in your hand?
Also I can imagine people could fall asleep when the car is driving.
In this Volvo movie the driver is opening a note book and making notes and also is watching a movie.
Both Tesla and Volvo have auto pilot mode which still needs an alert driver. But they advertise it like autonomous driving.
Edit: Volvo will launch Level 4 autonomous driving indeed. So they are a step ahead of Tesla. But it's confusing because they also call it auto pilot: http://www.volvocars.com/au/about/innovations/intellisafe/au...