Florida is certainly well-known for its gators and crocodiles. The handy rule of thumb is don’t mess with those cranky carnivores.
In addition to those semiaquatic reptiles, there is also an abundance of sea turtles. Florida is said to be the home of the densest population of nesting Loggerhead Sea Turtles in the entire Western Hemisphere. I dare say that most of us have a rather positive viewpoint about turtles and generally consider them to be relatively harmless and (shall I suggest?) cute or endearing (they are portrayed that way in movies and TV shows).
But even sea turtles can get down and dirty. If you approach a sea turtle and appear threatening, please know that they can bite you quite fiercely. The strong jaws of sea turtles are legendary and not to be trifled with.
The good news is that you can usually readily outrun a turtle.
About the only way to get bitten by a turtle involves your persistence in poking at them. Rarely would a sea turtle slink up behind you and catch you entirely off-guard. That’s the type of ploy that is usually reserved for those sneaky alligators and dastardly crocs.
Okay, so we probably can all agree that getting harmed by a turtle is a low odds proposition. Normally, you would need to go out of your way, quite a bit, in order to antagonize a turtle and prod it into a seething and overt attack toward you.
Leave them be, and you’ll be scot-free of inherent danger.
Of course, if a turtle smashes through your car windshield at high speeds and strikes you within the confines of your car, all bets are off. In that case, you likely are going to get banged up.
Say what?
Yes, a recent news report out of Florida indicates that a turtle burst through the front windshield of a car that was cruising along on the popular Floridian highway known as Interstate 95. The turtle struck a female passenger in the head. This was the 71-year-old mother of the driver. You can imagine the shock and horror of the driver upon witnessing a turtle suddenly and unavoidably puncturing through the windshield and plowing into your mom.
Turns out that the driver summarily jammed on the brakes and brought the vehicle to a quick stop, frantically calling 911 for assistance. The arriving first responders took her mother to a nearby hospital and thankfully the cut she got above the eye did not do any demonstrative damage. She will be okay.
Firefighters that had responded to the emergency were able to determine that the turtle was still alive and pretty much unhurt. They gingerly removed the turtle from the car and released it into the woods. One supposes that the turtle has quite a story to tell, being able to explain to all the other turtles about the day that it perchance landed inside a moving car and lived to tell the tale.
I’ll give you a moment to ponder the wonderment of the news story.
You might also be thinking about how the turtle got into such a predicament, to begin with. This was not a flying turtle, though it did manage to get airborne. Did the turtle want to undertake this highly hazardous act? Maybe it is the revered daredevil amongst all the other turtles in the vicinity.
The guess by local authorities is that the turtle was crawling along on the active highway and got picked up by a passing semi-truck that then launched it from the massive wheels and into the air. We can logically assume that the two women in the car were somewhere behind the semi-truck and had the bad luck of being in the wrong place at the wrong time.
In case you are questioning whether a turtle could actually penetrate through a car windshield, this is a strong possibility and it indeed happened in this particular case. The hardness of the shell and the weight of the turtle, along with its shape, and being propelled into the air, would all be factors in being able to strike the windshield with tremendous force and effecting a puncturing aptitude. Keep in mind too that the car was traveling along at a high rate of speed.
The point is that we can reasonably conclude that the story is true and that in fact, a turtle can crash through a car windshield. There are numerous reported cases. Sometimes the turtle comes all the way into the vehicle, while on other occasions it becomes lodged in the windshield. At times, the turtle survives the calamity, and sadly there are instances whereby the turtle succumbs to the effort and perishes.
For this latest incident, we can count this as a happy story since the woman struck by the turtle is going to be okay and the turtle was fine and returned to its natural habitat. You can’t always expect a happy ending in real life, but this was such an occasion.
Should you constantly be on the alert as a driver that a turtle might appear out of nowhere and shatter your windshield while driving along and minding your own business?
I don’t think so.
Unless you live in an area that is massively populated with turtles, you aren’t obligated as a driver to be continually scanning the horizon for turtles that might be errantly on the highways and byways that you traverse. Airborne turtles that strike cars are the stuff of fanciful (yet real) stories and not a matter that needs to preoccupy your mind while at the steering wheel of a moving car.
That being said, the overall concept of roadway debris that can fly up at your car, well, that’s something you indeed should be constantly scanning for. Rather than a turtle, maybe the object might be a discarded piece of junk, or perhaps a hood ornament that fell off the front of a fancy car, etc. Those semi-trucks and large vehicles can easily grab up the most innocuous of debris and toss it into the air, becoming an unguided missile that can ruin your day.
Shifting gears, consider that the future of cars entails the advent of self-driving cars. Self-driving cars are going to be using AI-based driving systems and there won’t be a human driver at the wheel (see my extensive coverage of self-driving cars at this link here).
Here is an intriguing question: How would AI-based true self-driving cars cope with a hurtling turtle that smashed through the windshield while underway at high speeds on an open highway?
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 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 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 Flying Turtles
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.
One aspect to immediately discuss entails the fact that the AI involved in today’s AI driving systems is not sentient. In other words, the AI is altogether a collective of computer-based programming and algorithms, and most assuredly not able to reason in the same manner that humans can (see my explanation at this link here).
Why this added emphasis about the AI not being sentient?
Because I want to underscore that when discussing the role of the AI driving system, I am not ascribing human qualities to the AI. Please be aware that there is an ongoing and dangerous tendency these days to anthropomorphize AI. In essence, people are assigning human-like sentience to today’s AI, despite the undeniable and inarguable fact that no such AI exists as yet.
With that clarification, you can envision that the AI driving system won’t natively somehow “know” about those airborne turtles. This is an aspect that needs to be programmed as part of the hardware and software of the self-driving car.
Let’s dive into the myriad of aspects that come to play on this topic.
First, we can consider the matter of whether it is vital to detect that the threatening object is a turtle.
Allow me to explain.
The sensors of the self-driving car are intended to examine the environment surrounding the vehicle and provide data about whatever is able to be detected. There are video cameras that are streaming live videos of the driving scene. There are usually radar devices, LIDAR units, ultrasonic detectors, and so on, all of which are attempting to sense the nearby scenery.
This torrent of data is then examined by the AI-based driving system. For example, there is image processing software that is running on the onboard computer processors. The software is oftentimes based on today’s Machine Learning (ML) and Deep Learning (DL) algorithms. In brief, these are computational pattern matching techniques that try to identify what an object is.
Programmers and engineers have usually pre-trained the ML/DL. They might have used thousands of pictures of cars and trucks to train the ML/DL on what such vehicles look like. Thus, when the video images come pouring into the system, the pattern matching can identify that a car is up ahead and a truck is further ahead of the car.
The pattern matching is done for pretty much all discoverable objects in the driving scene (well, not necessarily, since this could be computationally expensive and exhaustive). Anyway, the idea is that the AI driving system will need to identify where other cars are, where pedestrians are, and the like. This requires mathematically examining the sensory data and trying to ferret out which objects are present.
In addition, the detection should also be trying to figure out the movement of the objects that are being identified. A lamppost at the side of the rode is likely stationary and fixed in place. A parked car at the edge of the roadway is perhaps stationary at this moment in time but could get underway suddenly. A crucial element of the self-driving car and the AI driving system entails being able to find objects, categorize them, and predict what those objects are doing and what they might do next (for my deep analysis on the too many objects at a time concern, see the link here).
All right, with that important preamble, we can return to the matter at hand.
Should the AI driving system try to ascertain whether a turtle is a turtle, assuming that a turtle is findable within the driving scene?
One argument is that of course, the AI driving system ought to be programmed to identify turtles.
Suppose that a turtle is crawling along the roadway and a self-driving car is approaching the beloved creature. We don’t want the self-driving car to simply drive over the top of the sweet turtle. If the AI has simply categorized the turtle as say a boxy-shaped object of no known configuration, this implies that the AI would not be able to factor into the equation that this is a possible living animal.
That might be a significant factor when making a driving decision.
A human driver might opt to roll over a fragment of broken furniture that is laying on the street. You might do so because trying to swerve away could be dicey and get you into a worse car accident. Sometimes the easiest approach involves driving over a roadway piece of debris. At the same time, a human driver might take some desperate attempts to avoid striking a breathing animal, and thus knowing whether an object is a turtle versus a piece of scattered furniture is a notable aspect.
Please be aware that there are more twists and turns involved.
As a driver, if an object is flying straight toward your windshield, does it matter whether you are able to figure out what exactly the object is?
In one sense, it is obviously important to try and ascertain the size, weight, shape, and trajectory of the object. This can be done without necessarily being able to identify the specific nature of the object. All you need to know at that moment is that it is either big or small, going to be bouncing off your windshield or smashing into it, and could land inside the vehicle.
I recall that one day I was driving past a golf course and an errant golf ball came over the top netting and smacked into my windshield while I was going at about 45 miles per hour. It happened very fast. I did not see the golf ball while it was in the air. All I managed to see was something that at the last moment struck my windshield. I then watched as the white-colored and ball-like object bounced off the now cracked windshield and landed onto the highway, bouncing merrily along.
In a somewhat after-the-fact Sherlock Holmes deductive logic, I reasoned that it was a golf ball. Perhaps it could have been some other similar object, though the preponderance of anecdotal evidence suggests it was likely an everyday golf ball. Luckily, other than the shocking impact, the windshield remained intact, I could still see the road, and I kept driving. Later on, I had to replace the severely cracked windshield and was cursing those idiot golfers that couldn’t keep their game on the golf course (I don’t like to use foul words, so I won’t refer to the golfer as a duffer).
Returning to the flying turtle, in the case of a turtle that was hurled into the air and is stridently heading toward the windshield of a self-driving car, it would seem arguable whether or not the AI should necessarily need to be programmed to identify this darting object as a turtle (i.e., this ostensibly is a nice to have capability versus an ironclad required function). You could readily assert that as long as the AI driving system is able to assess the size, shape, and other factors, the act of not discerning that it is a turtle could be considered adequate in this particular scenario.
I bring this up for several reasons.
One reason involves the so-called edge or corner case problem that confronts the advent of self-driving cars.
An acrimonious issue being hotly debated consists of the zillions of edge or corner cases that can arise when driving a car. Some believe that self-driving cars ought to not even be on our public roadways until all possible scenarios have been tested via a simulation or via a closed track testing facility.
Others point out that it would likely never be the case that self-driving cars would ever get onto our roadways if you first had to figure out every conceivable scenario. There are always additional variants that can be dreamed up and thus the number of possible permutations would seemingly be endless.
As a result, most of the automakers and self-driving car tech firms are concentrating on the more likely scenarios and leaving the edge or corner cases for being dealt with later on. This would seem like a prudent and realistic way to get underway with self-driving cars, especially if you believe that a prevalence of self-driving cars will ultimately reduce the number of car crashes and consequently reduce the number of human injuries and fatalities associated with driving.
For those that accept the notion of proceeding and then, later on, coping with edge cases, there is another problem at hand, namely, what constitutes a reasonable definition for an edge case? Suppose a self-driving car tech firm decides that jaywalking pedestrians is an edge case, and it can be dealt with later on. I doubt that any reasonable person would decidedly go along with that scenario as an edge case and instead would insist it needs to be a core or cornerstone capability of the AI driving system.
Time for a brief sidenote. That was notably a somewhat outrageous or outlandish exemplar and so don’t be dinging me about the idea that anyone in their right mind would claim that jaywalking pedestrians is an edge case. I was trying to use some provocative hyperbole to help make a point. Please go with the flow.
The totality of all of this back-and-forth is that the use case of the hurtled turtle could be “reasonably” handled as any generic kind of debris that is coming at the windshield of the self-driving car. Having to figure out in real-time that it is a turtle is not quite as vital as is ascertaining the profile of the object and what it might do.
Conclusion
The saga of how a self-driving car would deal with a flying turtle has only just begun.
There is a lot more to be considered, though due to space constraints I can’t cover the whole shebang.
In the case of the Florida driver, the driver likely had little if any recognition of the turtle before it struck the windshield. By this, I mean that the driver probably not only did not realize it was a turtle, but the driver might also not have consciously been aware of the object at all (similar to how I was taken by surprise by the alleged golf ball).
Ergo, a human driver can get caught unawares.
For an AI driving system, there is a chance of that kind of being caught unawares, though only likely in the sense that the object might be moving so fast that there was insufficient time for the computer processors to do the needed image analyses and electronically calculate what to do.
To clarify, it would seem generally likely that the sensory devices captured imagery or other data that would record the aspect of a flying object like this (assuming that it was visible and otherwise not obscured or hidden). The problem would next consist of how fast the hardware and software could undertake the needed analysis.
Even if the computational processing was done within sufficient time, the next question is what should the AI driving system do about the emerging predicament. Should the self-driving car swerve? Should the self-driving car go faster and try to get under the object, maybe having the object hit the roof of the car instead of the windshield? Should the self-driving car try to hit the brakes? And so on.
These are all very hard questions and bring to the foreground the famous (or some say infamous) Trolley Problem (see my coverage in my columns).
Suppose the turtle hits the windshield, enters violently into the vehicle, and strikes a passenger. This could absolutely befall a self-driving car. The question then arises as to whether the AI driving system will have detected this infiltration and be able to ascertain what driving action is next most suitable to perform. Hitting the brakes might not be a proper action, since doing so could cause cars behind the struck car to get hit from behind.
All told, this turtle story provides a plethora of handy points about what self-driving cars need to be programmed to deal with.
Fortunately, the humans were okay, and the turtle did not lose its life while garnering this opportunity to glean some lessons learned. One hopes that the turtle though reads this article and realizes that meandering into the highway is a risky gambit, including even once self-driving cars are abundantly on our highways and byways.
Stay safe, courageous turtle.
