Considering the predicament of self-driving car getting entrapped in a snowbank.
getty
Snowstorms aplenty are upon us as the winter decides it is time to let loose.
A recent news report about a snowfall calamity that luckily had a happy ending is both surprising, an eyecatcher for sure, and helpful to ponder and mindfully reflect upon.
Here’s the story.
A man in upstate New York inadvertently drove off an icy roadway, plowed into a bank of snow, and while inside the car, he became entrapped when an abundance of accumulating snow engulfed his vehicle. As if that wasn’t already unnerving enough, apparently snowplows might have shoveled snowy buildup onto his trapped automobile when attempting to clear the adjacent roadway of the rapidly falling snow (reportedly, the local area got about four feet of snow).
Let’s do a quick recap.
Imagine yourself stuck inside a car, buried under by snow, and having no ready means of escape. It seems like one of those wild stunts that a Houdini-inspired prankster might try to pull.
But this was deadly serious business.
Freezing cold temperatures were underway. He had gone off-the-road at about midnight. It was dark with arctic-like weather conditions. Inside the car, the bitter cold was taking its toll on the man. His body temperature began to wane, and the looming possibility of hypothermia and frostbite became a frightening reality.
I’m sure that we all have had our own worries that we might one day be driving and do a slip-and-slide off an icy and snowy road, but the worst-case scenario was that your car might get wrecked and you’d need to hoof your way to the nearest point of civilization. In this case, he was in the midst of civilization, wherein his car was embedded into a snowy embankment that was amid a series of homes and mailboxes. In other words, this didn’t happen in some remote off-the-beaten-path and instead occurred in the middle of a relatively populated area.
It seems impossible to envisage that you could be buried inside a snow-covered car while nearby are people inside their homes, sitting in front of cozy fires in their fireplaces, perhaps lazily looking out their windows at the fluttering snowflakes and enjoying steaming mugs of hot cocoa.
As they say, reality is sometimes stranger than fiction.
Well, your first thought might be that he must have had a cell phone, since everyone has a cell phone nowadays, and all he would need to do is call 911 to have someone come and rescue him. According to the media reports, he did try calling 911, repeatedly, but his cell phone reception was extremely weak, and he was, unfortunately, unable to get a solid signal. As a side note, I suppose that if you’ve ever cursed the phone companies about their cell coverage, this is an instance whereby you really would have a bona fide beef about it.
Eventually, around 8:30 a.m. in the morning, his cell phone managed to finally get sufficiently connected with 911. This good luck was undercut by the bad luck that the geolocation of where he was had troubles being determined, and the best guess was that he probably was somewhere along a 3-mile stretch of roadway. Presumably, he either did not know where he exactly was or might have been unable to convey his position via the spotty cell phone connection.
The local police drove along the expected roadway to search for the snow-covered car. One policeman stated that he stopped at numerous snowbanks to dig into the snow to see if the vehicle was buried within. Finally, after the search had been underway for quite a while, and at around 10 a.m. or so, the policeman that was periodically digging into snowbanks noticed a large clump of snow that was in front of a house. It was not considered a suspected place for the car to be.
Thinking that the snow was covering mailboxes, the cop opted to plunge his arm deep into the snow to clear a hole to spot the address on the mailbox. Turns out, he startlingly realized that he had just punched the side of a car window.
The policeman was standing atop the snow buried car.
With the help of a passerby, the policeman pulled the nearly frozen man out of the car and then loaded him into the police car. The icy man was taken quickly to a nearby ambulance and then transported to a local hospital. The man managed to survive this oddball “snow buried car” dilemma and reportedly was doing okay at the hospital.
That’s the gist of the story, now here’s some after-the-fact couch-based quarterbacking going on.
Some have already criticized the man for not having snow-related gear in his car such as perhaps having the kind of stuff you’d wear when going skiing or camping in the snow. Sure, we all ought to be always prepared, but it seems a bit farfetched and one of those hindsight-is-always-right exhortations to unduly ding him for not being packed to the gills with cold-weather gear.
Note to self: Go right now and put my winterized insulated goose-down filled sleeping bag in my car as a future scenario precaution.
Another criticism in our smarmy world of today is that he should have broken a car window and dug his way out.
First, you’d need some viable means to break out one of the car windows, which it turns out are a lot more resilient than you might assume. The retort there is that we should all be riding around with one of those specialized tools that can smack and smash a car window.
Second, if you did somehow get through a window, what would you use to dig or tunnel your way out of the snow cave that you are immersed in? The flippant answer is that you ought to be carrying a hiker’s collapsible shovel with you at all times, presumably in the trunk of your car (which, you’d need to stridently reach into from the backseat of the car, assuming this was feasible to undertake).
Or, all in all, you might do a MacGyver, perhaps by tearing apart a car seat, and then cleverly using the various metal and plastic components as both a window smasher and then as a digger to burrow your way through the snow (for more about being a modern-day MacGyver and cars, see my coverage at this link here).
It is stunningly amazing the number of snippy comments that people can come up with on these grim occasions. Some say he shouldn’t have driven off-the-road, to begin with, and implied he deserved whatever happened next. Really? That’s a pretty tough outlook on life. Others wondered why he didn’t just run the car engine and use the car heater to keep warm. Reportedly, the car was no longer functioning after going off the roadway.
Yes, we can come up with dozens of questions and zillions of variations. One supposes that if nothing else, this tale of survival maybe gets people thinking about the topic overall and might prepare any of us for the day that some similar untoward scenario happens in our daily lives.
Speaking of thinking contemplatively about these dire situations, consider that the future of cars is that we will inevitably be using self-driving cars.
Here’s an interesting question: What would happen if an AI-based true self-driving car were to slide off a road and end up in a snowy embankment?
Let’s unpack the matter and see.
Understanding The Levels Of Self-Driving Cars
As a clarification, true self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.
These driverless vehicles are considered a Level 4 and Level 5 (see my explanation at this link here), while a car that requires a human driver to co-share the driving effort is usually considered at a Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).
There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.
Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some contend, see my coverage at this link here).
Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional vehicles, so there’s not much new per se to cover about them on this topic (though, as you’ll see in a moment, the points next made are generally applicable).
For semi-autonomous cars, it is important that the public needs to be forewarned about a disturbing aspect that’s been arising lately, namely that despite those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.
You are the responsible party for the driving actions of the vehicle, regardless of how much automation might be tossed into a Level 2 or Level 3.
Self-Driving Cars And Being Snowbound
For Level 4 and Level 5 true self-driving vehicles, there won’t be a human driver involved in the driving task.
All occupants will be passengers.
The AI is doing the driving
This means that if there is a passenger inside the self-driving car, they would be innocent of driving the car off the roadway, and thus those that were earlier tossing acrid barbs at the man that drove off the road in the snowy conditions could no longer blame a human driver for such a transgression.
In turn, this might get you wondering whether the AI would somehow inadvertently drive off the road.
Some seem to think that AI driving systems will always be driving in a perfect manner. By perfection, they seem to suggest that the car will defy the laws of physics and otherwise be acting as though operating in a fairytale world.
Let’s be clear about this, a car is a car. And, the AI is not omniscient, nor able to avoid all possibilities of calamity and roadway problems.
We are going to have car accidents in an era of self-driving cars.
Forget those hollow slogans that say we are going to have zero fatalities, which I’ve repeatedly proffered has a zero chance of happening (see my discussion at this link here). The hope is that we’ll have a lot less crashing and colliding of cars, which I agree should so occur, but it will not magically eliminate all car-related fatalities and injuries (it’s a big number right now, amounting to about 40,000 human fatalities per year and approximately 2.3 million injuries, that’s only in the United States and the numbers are much higher on a global basis).
In short, the AI driving system could indeed end up in a predicament such that the AI was unable to keep the self-driving car from sliding off the roadway. You could in fact have a self-driving car that goes into a snowy embankment.
It can happen.
It will happen.
Envision that the self-driving car is making its way in snowy and icy conditions. The roadway is super slick. Despite all the state-of-the-art sensors, including video cameras, radar, LIDAR, thermal imaging, ultrasonic, and so on, the roadway status can still be murky and not fully detectable. The car comes upon an especially bad patch, and the wheels and the vehicle begin to skid.
What will the AI do?
I’ve previously described that some of the AI driving systems are being infused with specialized driving capabilities for these snowy and icy buildup driving situations (see my analysis at this link here). In that case, the AI would invoke the driving skillset for this particular circumstance. That will certainly help but still does not provide a guarantee that the vehicle won’t nonetheless go off the road.
If you’ve been following along about the progress of self-driving cars, you might have noticed that most of the self-driving car experiments on our public streets are being done in warm weather climates, such as California, Arizona, and Florida. This is partly because the driving efforts in those locales don’t have an abundance of snow and seldom have ice, thus they are a lot easier to deal with. In that sense, many of the existing self-driving cars are not yet fully prepared to drive in those freezing cold temperatures that embody adverse roadway conditions.
In the self-driving car industry, there is specialized parlance that refers to the Operational Design Domain (ODD) for a vehicle. The ODD is an indication of the stipulated conditions that the AI driving system can handle.
For example, an automaker and self-driving tech firm might specify that their existing AI driving system can cope with warm weather and sunny days but is not yet capable of driving in cold weather and darkness. Meanwhile, some other AI driving system by another maker might be trained for nighttime driving and daytime driving, but only is capable in city environs and not ready for outlier locations. And so on.
Having an ODD that encompasses the extremes of driving situations, such as driving in snowy conditions, generally is considered a lower priority and placed onto the edge or corner case list. This means that this type of driving will be dealt with later on. Right now, the belief overall is that it is timelier and more sensible to focus on day-to-day driving in the city and suburbs, which is where the money is most likely to be made by providing a so-called robo-taxi service.
The AI driving system is supposed to figure out when its ODD is being exceeded.
Let’s say that you get into a self-driving car and it starts on a journey to your grandmother’s house. During the driving trip, the weather goes afoul. If this bad weather is considered outside the scope of the ODD, the AI driving system is intended to detect this aspect and then gradually and safely bring the car to a halt.
I’ve pointed out that this will be quite disturbing for people that are riding in self-driving cars. You might be inside a self-driving car, and if it had a human driver the vehicle could have kept going, but since the AI is in-charge, and the vehicle has reached its ODD boundaries, the AI will refuse to continue driving and park you on the side of the roadway or in some such considered “safe” place.
Before you start to carp about this functionality, it does bring up an interesting point related to the story of the man that went into the snowbank.
If AI had been driving the car, perhaps it might have earlier determined that the roadway conditions were overly dangerous, or the conditions were outside the defined ODD for that AI driving system, and ergo it would have not kept driving. By not driving, it would have presumably avoided sliding off the roadway.
Problem solved by never getting into the problem per se. Of course, that was one of the unctuous comments that some had already uttered, namely blaming the driver for driving in conditions that he should (apparently) not have been driving in.
This is somewhat easy to assert as a broad statement, but the devil is in the details.
Suppose that you are heading to your grandmother’s house because you have her pills and she is in bad health, desperately needing her latest batch of medicine. You probably are willing to take a heightened chance at driving in adverse conditions. We all make such decisions. When it comes to an AI-based self-driving car, you as a passenger will essentially no longer be in the driver’s seat and therefore cannot make those kinds of tough decisions.
It will be quite fascinating to see how this works out.
People will presumably no longer be able to make high-risk driving journeys, at least with respect to whatever ODD falls within that high-risk category. On the other hand, even for moderate risk trips, the AI is presumably going to be a better driver, ostensibly because it won’t drink and drive, nor will it drive distracted, and therefore we ought to have fewer fatalities accordingly.
On the balance, how will society perceive the benefits of self-driving cars versus the “costs” associated with perhaps not being able to use a car whenever you wish to do so?
This might also be a basis for some people clinging to keeping human-driven cars around and argued as a basis to avert any movement toward having only self-driving cars on our roadways. There is a contingent that insists we need to eventually and inexorably get rid of all human driving, though this is quite a radical proposition and opponents often say that you will only get them out of the driver’s seat whence you have pried their cold dead hands from the steering wheel.
Conclusion
The point of the aforementioned context about self-driving cars is that there will still be a possibility of even a self-driving car going off the roadway and into a snowy embankment.
There’s more to consider on this matter.
It is assumed that by-and-large self-driving cars will be deployed in fleets by large companies on a ridesharing basis.
As such, those fleet operators are going to be keeping close tabs on where their self-driving cars are at all times (doing so to know when the vehicle is available for its next money-making ridesharing requested ride). The odds are that if a self-driving car did slide into a snowbank, the fleet operator would likely know where this happened, and be aware that the vehicle had stopped moving. In that case, presumably, the fleet would send out a rescue mission or contact 911, right away.
Another facet of self-driving cars is their OTA (Over-The-Air) electronic communications capabilities. This allows the fleet operator to push updates into the AI software and allows the uploading of the driving data from the vehicle. It is likely that if an AI driving system got itself into a bad spot, it would do the classic “phone home” to alert the fleet operator.
There is also the use of V2V (vehicle-to-vehicle) electronic communication.
We are going to have self-driving cars that convey electronic messages to each other. A self-driving car that detects debris on the roadway could alert any nearby self-driving cars to be watchful. In that same manner, a self-driving car that has encountered a slick roadway could warn other self-driving cars. And, if a self-driving car perchance got into a pile of snow, it could try using the V2V to send an alarm to other nearby self-driving cars, which they could then take appropriate rescue actions.
Here’s a final twist that you might find poignant.
The existing standard about the levels of self-driving cars does not cover the aspects of off-road driving.
In essence, there isn’t an explicit requirement to be able to drive in an off-roading situation. Some automakers aren’t dealing with the off-road facets as yet, since it seems apparent that making money from self-driving cars is first going to come from driving on-roads rather than when off-road in deserts, pastures, the mountains, or similar.
Once an AI self-driving car ends up in a snowbank, it would presumably not have the capability included to cope with trying to drive out of it.
Of course, a human driver might not be able to do so either.
Eventually, the odds are that AI driving systems will include off-roading provisions, which will be somewhat curious to see. I say this because some believe that the fun of going off-roading is being at the wheel and having to contend with the challenges of the driving experience. Will we equally enjoy an off-roading jaunt when the AI is doing all the driving?
Anyway, please be watchful of those snowy and icy roads that are increasingly appearing as the winter embrace strengthens, and if you see a car completely covered in snow, I guess you cannot assume it is empty. There might be a human in there.
In the future, there might not be a human at all in the vehicle, but there might be AI, although the AI won’t be worried about the snowy entrapment, other than that it ostensibly wants to perform its duty of driving a car and cannot do so until the vehicle is released from its snowed-in confinement. Help me, the AI might be decrying, help me to help you, by getting back onto the roadway and driving people wherever they want to go.
Well, within reasonable limits, one might say.
